User-Generated Censorship:Manipulating The Maps Of Social Media

by

Christopher E. Peterson

B.A., Legal StudiesUniversity of Massachusetts at Amherst, 2009

SUBMITTED TO THE PROGRAM IN COMPARATIVE MEDIA STUDIES/WRITING IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF

MASTER OF SCIENCE IN COMPARATIVE MEDIA STUDIESAT THE

MASSACHUSETTS INSTITUTE OF TECHNOLOGY

JUNE 2013

© 2013 Christopher E. Peterson. Some rights reserved. This work is licensed under the Creative Commons Attribution-NonCommercial-ShareAlike

3.0 License (CC BY-NC-SA 3.0).

The author hereby grants to MIT permission to reproduce and to distribute publicly paper and electronic copies of this thesis document in whole or in part

in any medium now known or hereafter created.

Signature of Author: _____________________________________________________________Program in Comparative Media Studies/Writing

May 24, 2013

Certified by: ___________________________________________________________________Ian Condry

Associate Professor, Comparative Media StudiesHead, Foreign Languages and Literatures

Thesis Supervisor

Accepted by: ___________________________________________________________________Heather Hendershot

Professor of Film and MediaDirector, Comparative Media Studies Graduate Program

User-Generated Censorship:Manipulating The Maps Of Social Media

by

Christopher E. Peterson

Submitted to the Program in Comparative Media Studies / Writing on May 24th, 2013, in Partial Fulfillment of the

Requirements for the Degree of Master of Science in Comparative Media Studies

Abstract:

The last decade has seen the rise of new technologies for making information more broadly available and accessible. Variously called ‘user-generated content,’ ‘social media,’ ‘social news,’ ‘crowd-curation,’ and so on, these design conventions, algorithmic arrangements, and user practices have been widely praised for ‘democratizing’ media by lowering the barriers to publishing, accrediting, and aggregating information. Intermediary platforms like Facebook, reddit, and Twitter, among others, are generally expected to elicit valuable knowledge through the algorithmic filtering mechanisms broadly distributed among their users.

This thesis investigates user-generated censorship: an emergent mode of intervention by which users strategically manipulate social media to suppress speech. It shows that the tools designed to help make information more available have been repurposed and reversed to make it less available. Case studies reveal that these platforms, far from being neutral pipes through which information merely travels, are in fact contingent sociotechnical systems upon and through which users effect their politics through the power of algorithms. By strategically pulling the levers which make links to sites more or less visible, users recompose the representations of the world produced by social media, altering pathways of access and availability and changing the flow of information.

This thesis incorporates insights from media studies, sociology, law and policy, information science, and science-technology studies to study user-generated censorship. It contributes to a broader conversation now emerging across fields which seeks to explore and understand the politics of our developing social media systems.

Thesis Supervisor: Ian CondryTitle: Associate Professor, Comparative Media Studies

2

PREFACE

Biographical Note 4

Acknowledgements 5

I: INTRODUCTION

Suppressing J30Strike 6

Questions, Definitions, Cases 8

Outline 12

II: A CASE STUDY IN USER-GENERATED CENSORSHIP

“Bury Until They Change Their Ways”: Introducing the Digg Patriots 14

“Any Chair In A Bar Fight”: Mission, Metrics, & Methods 18

“A Rude Wake-Up Call”: Reading the Digg Patriots 28

III: UNDERSTANDING USER-GENERATED CENSORSHIP

As Corrupting Collective Intelligence 31

As Reintroducing Heteronomy to the Networked Public Sphere 36

Or: As Actor-Networks Assembling Artifacts 42

IV: THE SOCIAL PRODUCTION OF WHAT?

The Ark in the Archives: On the Centrality of Sociotechnical Systems 46

Social Media as Machines which Manage the Flood 49

The Social Production of Persuasive Panoramas 55

V: CONCLUSION

The Relevance of User-Generated Censorship 59

APPENDIX

A.1: The Digg Patriots Data: Origins, Controversies, Considerations 66

A.2: The Digg Patriots Data: Some Exploratory Visualizations 70

3

PREFACE

Biographical Note

Chris Peterson graduated from the University of Massachusetts, Amherst in May 2009, with B.A. in Critical Legal Studies and a minor in Information Technology. He spent six semesters as a teaching assistant for Law and the World Wide Web, Law in Virtual Worlds, and Legal Fictions. His senior thesis, supervised by Professors Ethan Katsh and Alan Gaitenby, adapted approaches from law, sociology, design, behavioral economics, and human-computer interaction to study Facebook privacy, and was featured in the New York Times.

Peterson has worked as a senior campus representative for Apple; on the bargaining committee of United Auto Workers 2322; as a research assistant at the Berkman Center for Internet and Society; and most recently as an admissions officer at MIT, where he directed digital strategy while also participating in the selection process with a special interest in subaltern and maker applicants. As a graduate student, he was a research assistant at the Center for Civic Media in the MIT Media Lab, where he managed undergraduate researchers and oversaw data journalism projects. He serves on the Board of Directors at the National Coalition Against Censorship and as a Fellow at the National Center for Technology and Dispute Resolution.

Barring an extinction event, he can probably be reached via email at chris [dot] peterson [at] mit [dot] edu and/or via the web at http://cpeterson.org.

4

Acknowledgements

Writing a thesis is hard; writing a Comparative Media Studies thesis doubly so. The great reward of a CMS education is that it imparts a penetrating perspective on the biases of media broadly defined, including of the very tools (technological, methodological, ideological) used to study media. The associated risk, however, is that the tools of inquiry themselves may suddenly transform from ready-to-hand to unwieldy instruments, which must themselves be wrestled while the intended subject of study threatens to escape cackling into the distance.

I captured it only with the aid and support of my committee. Ian Condry agreed to be my thesis advisor after we’d barely met; our long talks, and his frank advice, led me through the labyrinth this last year. Ethan Zuckerman, despite being the most brutally busy person I know, made the time to read my drafts and provide detailed feedback. Nancy Baym gave freely of her qualitative expertise and insightful observations. I am very grateful to them.

I received helpful nudges from many others along the way. My friend and colleague Drew Harry, who let me frequently pick his considerable brain, was a committee member in all but name. Beth Hadley ’15, Betsy Riley ’14, and Steve Murray helped me write (and fix) my code. Ann Blair, Finn Brunton, Susan Crawford, Charlie DeTar, Peter Galison, Stuart Geiger, Eric Gilbert, Tarleton Gillespie, James Grimmelmann, Del Harvey, Max Kantor, Markus Krajewski, Rebecca MacKinnon, Nick Montfort, Evgeny Morozov, Jeff Ravel, Clay Shirky, Dave Weinberger, and Jonathan Zittrain are just a few of the scholars, students, and professionals who volunteered their time and thoughts as I conducted my research.

The staff helped me stay sane through the slog. Becky Shepardson, Jess Tatlock, and Andrew Whitacre of CMS; Shannon Larkin, our tireless graduate guardian angel; and Lorrie LeJeune of Civic, who dispensed chocolates, tea, plants, and career advice with calm and care. Jim Paradis, who first encouraged me to apply to CMS, is a rock for the entire program.

My classmates have been amazing. I have learned more from their brilliance, their character, and their kindness than I could have imagined. So too my friends in Civic, who always knew when and how to deliver exactly the appropriate mixture of snark and support.

I would not be here except for the MIT Office of Admissions, which first welcomed me into this strange, spectacular place and gave me the freedom to do good work. Nor would I still be here without the undergraduates I have come to meet, advise, and befriend. They are among the most inspiring, incredible people I have ever known. Sam Keyser once quipped that MIT is hard to get into, and even harder to leave, and it is these students who make it so. Among them, I often feel I have finally found my people; IHTFP.

I have been blessed with an incredible family. This thesis is dedicated to them, and especially to my grandfather Gorham. Though not an educated man, he has always had a love of learning, some of which seems to have rubbed off on me. Much of my childhood consisted of him patiently telling me that there was nothing new under the sun and to always pay close attention to the text. He would have felt at home in many of my classes at CMS, and I think I came to feel at home here in no small part because of him.

5

I: INTRODUCTION

It was the spread of the message that mattered [in the history of communication] — not its origin but its amplification, the way it reached

the public and ultimately took hold. That process should be understood as a matter of feedback and convergence, rather than one of trickling down

and linear causality.- Robert Darnton, “News and the Media in Eighteenth-Century Paris”1

1. Suppressing J30Strike

In June 2011, as heat and hardship beat down on Britain, progressive activists

proposed a general strike to protest austerity measures. Like many movements over the last

decade they began organizing the strike online.2 They created a website at J30Strike.org and

launched a publicity campaign through social media, especially by sharing links to the site

through Facebook’s News Feed.

It’s easy to understand why. Facebook is a top source of visitors for the five most

popular news sites in the United States.3 The analytics company Shareaholic estimates that

Facebook refers more than 25% of global traffic it tracks.4 Facebook’s News Feed does more

than just capture and redistribute eyeballs: like the front page of a major newspaper, it also

articulates an agenda, assembling a summary digest of important events. “As more and more is

shared,” wrote engineer Peter Deng after Facebook centrally embedded the News Feed, “we

want you to be able to find out everything that is going on in the world around you.”5 It’s a

6

1 Robert Darnton,“An Early Information Society: News and the Media in Eighteenth-Century Paris.” The American Historical Review 105, no. 1 (February 2000), available online at http://www.jstor.org/stable/2652433 and http://www.historians.org/info/aha_history/rdarnton.htm/.2 See generally Clay Shirky, Here Comes Everybody: The Power of Organizing Without Organizations (New York: Penguin Press, 2008). For a critique of the role of networked technologies to in supporting social movements, compare Evgeny Morozov, The Net Delusion: The Dark Side of Internet Freedom (PublicAffairs, 2012). 3 Ross Dawson, “Where News Website Traffic Comes from: Google Vs Facebook - Trends in the Living Networks.” Trends in the Living Networks, May 9, 2011, http://rossdawsonblog.com/weblog/archives/2011/05/where-news-website-traffic-comes-from-google-vs-facebook.html/.4 Colleen Taylor, “Study: Pinterest Drives More Referral Traffic Than Google+, Nearly on Par with Twitter.” GigaOM, January 31, 2012, http://gigaom.com/2012/01/31/pinterest-referral-traffic-google-plus-twitter/.5 Peter Deng, “Welcome to Your New Home Page.” The Facebook Blog, March 11, 2009, http://blog.facebook.com/blog.php?post=59195087130/.

vision of social media as a kind of map, as an atlas informing users of worthwhile destinations

and providing routes, in the form of links, through which they may be reached.

But, ten days before the strike, Facebook began notifying the activists that links to

J30Strike.org could not be shared because they “[contained] blocked content that has

previously been flagged as abusive or spammy” by other users.6 It erased all links to J30Strike.

Then, with relentless, recursive efficiency, Facebook blocked links to sites which themselves

linked to J30Strike, including blog posts informing other activists of the embargo.7 J30Strike

suddenly vanished from the picture of the world projected by the News Feed. It wasn’t filtered

by a government or corporation. Its servers weren’t disabled by hackers. J30Strike was still

perfectly accessible but had become strangely unavailable. Like a rural village erased from a map

of the English countryside but not from the countryside itself, the site was still there, but

suddenly became much less likely to be found by casual travelers.

I knew some of the J30Strike activists. I wrote one of the blog posts which was blocked

by Facebook.8 Watching J30Strike disappear from my map disturbed me. My News Feed had

indeed appeared a comprehensive record of everything important going on in the world around

me, but my sudden inability to link to J30Strike destabilized this perspective, revealing instead

its highly contingent character. What I and others were allowed to see depended upon a

complex and invisible confluence of forces largely beyond our control. I began to wonder: what

else was being hidden from me? How? And by whom?

7

6 A copy of the message can be seen at http://etc.cpeterson.org/MIT/cms/thesis/images/j30strikeblocked.jpg/.7 See, e.g., my blog post: http://www.cpeterson.org/2011/06/20/facebook-censors-citizen-activism-website/, which was blocked with this message: http://etc.cpeterson.org/MIT/cms/thesis/images/blogpostblocked.jpg/.8 For more on J30Strike, see Nick Baumann, “Why Did Facebook Block UK Strike Site?” Mother Jones, http://www.motherjones.com/mojo/2011/06/why-did-facebook-block-uk-strike-website/.

2. Questions, Definitions, Cases

These questions have a negative valence, but the purpose of this thesis is not to argue

whether the Internet is ‘good’ or ‘bad’ (or, as Evgeny Morozov might tartly observe, whether

“the Internet” is a coherent category of inquiry itself).9 Instead, its purpose is to explore some

questions prompted by my experience with J30Strike. These include: how are the pictures of

the world produced by social media projected? By whom? With what effects? For what

reasons? More specifically, I aim to investigate an emergent practice which I call user-generated

censorship: an evolving mode of intervention by which users strategically manipulate social

media to suppress information.

Countless scholars have studied what is commonly called ‘Internet censorship,’

including the ‘intermediary censorship’ of social media platforms.10 Others, most prominently

Eli Pariser and Cass Sunstein, have argued that the feedback loops produced by unwitting

users interacting with algorithms may cause common lifeworlds to diverge.11 User-generated

censorship, however, differs from these more familiar models. J30Strike did not disappear

from my News Feed because it was targeted by some devious autocrat within a corporation or

state, nor because I signaled a personal disinterest in radical politics. Instead, it was made to drop

by other users, whose decentralized signals were aggregated and enforced through a

centralized platform, creating a new capacity for many distributed actors to suppress speech

through the power of the algorithm.

8

9 Evgeny Morozov, To Save Everything, Click Here: The Folly of Technological Solutionism (Public Affairs, 2013). 10 See, e.g., Access Denied: The Practice and Policy of Global Internet Filtering (Cambridge, Mass: MIT Press, 2008); Access Controlled: The Shaping of Power, Rights, and Rule in Cyberspace (Cambridge, Mass: MIT Press, 2010), particularly chapter 5 by Ethan Zuckerman, available at http://www.access-controlled.net/wp-content/PDFs/chapter-5.pdf); Rebecca MacKinnon, Consent of the Networked: The World-wide Struggle for Internet Freedom (New York: Basic Books, 2013). 11 Eli Pariser, The Filter Bubble: How the New Personalized Web Is Changing What We Read and How We Think (Penguin Books, 2012); Cass Sunstein, Republic.com 2.0 (Princeton, N.J.; Woodstock: Princeton University Press, 2009).

Jillian York of the Electronic Frontier Foundation has documented dozens of

grassroots campaigns to remove content from Facebook.12 In one well-documented case, Brian

Ries of the TheDailyBeast “decided to conduct a little experiment,” campaigning on his blog to

have followers mark a post by Sarah Palin as “racist/hate speech.”13 According to Facebook the

post did not violate any of its content standards, yet within 24 hours the post had nonetheless

been removed by what spokesperson Andrew Noyes called “an automated system” driven by

Ries’ actions.14 As Ries later observed, “it took the coordinated actions of a community a

fraction of [Facebook’s] size” to change the content available through it.15

Another example occurred on reddit, a popular ‘social news’ site where users vote on

links which an algorithm then organizes into ranked lists. In early 2012 some critics of

Republican Presidential candidate Ron Paul discovered that anything they posted, anywhere

on reddit, was being downvoted dozens of times within minutes. It was later revealed that a

libertarian redditor had written a program which allowed any Paul supporter to enroll her

account in a botnet which stalked these critics and rapidly downvoted anything they posted.16

LibertyBot, as it came to be called, effectively enforced diminished visibility on its enemies,

sinking their content and comments so far down that other users would be significantly less

likely to see them.

9

12 Jillian York, “Policing Content in the Quasi-Public Sphere.” OpenNet Initiative, September 2010. https://opennet.net/sites/opennet.net/files/PolicingContent.pdf; Jillian York, “AllFacebook.com Editor Uses Bully Pulpit in Attempt to Remove Facebook Page,” March 21, 2011, http://jilliancyork.com/2011/03/21/allfacebook-com-editor-uses-bully-pulpit-to-remove-facebook-page/.13 Brian Ries, “Help Report Sarah Palin’s Ground Zero...” http://moneyries.tumblr.com/post/841249526/tumblr-help-report-sarah-palins-ground-zero/.14 “Facebook Apologizes for Deletion of Palin Post.” CNN, July 22, 2010, http://politicalticker.blogs.cnn.com/2010/07/22/facebook-apologizes-for-deletion-of-palin-post/.15 Brian Ries, “My Facebook War With Palin.” The Daily Beast, July 23, 2010, http://www.thedailybeast.com/articles/2010/07/23/palins-facebook-ground-zero-mosque-post-how-it-disappeared.html/.16 Chris Peterson, “The Fault, Dear Reddit, Is Not In Our Bots, But In Ourselves: The Case of LibertyBot.” MIT Center for Civic Media, March 22, 2013, http://civic.mit.edu/blog/petey/the-fault-is-not-in-our-bots-dear-reddit-but-in-ourselves-the-case-of-libertybot/.

Google’s PageRank algorithm famously interprets incoming links as a kind of vote for a

website such that, all else held equal, more links translates to a higher ranking.17 For many

years marketers have worked to manipulate search results by writing programs which post

millions of links pointing to clients’ sites in order to increase their PageRank: one 2006 study

by Kolari et al. found that nearly 90% of all English-language blogs are in fact spam blogs (or

“splogs”) created to game search engines.18 Google has long forbidden the practice, and, in

April 2012, began penalizing sites with large numbers of “unnatural links” leading to them in

order to disincentive splogs.19 In response some smart spammers simply flipped their business

models, launching “NegativeSEO” companies which offered clients the ability to point billions

of “unnatural links” at their competitors to sink them in search.20

All of these platforms are influential intermediaries which incorporate user feedback to

help decide how prominent the pathways to a particular website should be through them, and

the ‘democratizing’, ‘participatory’ rhetoric has given rise to associated assumptions that the

arrangement of these pathways can be trusted by default. As Tarleton Gillespie has observed,

we seem to want the algorithms of social media “to be neutral, we want them to be reliable, we

want them to be the effective ways in which we come to know what is most important.”21

10

17 Page et al. “The PageRank Citation Ranking: Bringing Order to the Web.” (1999). http://ilpubs.stanford.edu:8090/422; Brin & Page. “The Anatomy of a Large-scale Hypertextual Web Search Engine.” Computer Networks and ISDN Systems 30, no. 1 (1998): 107–117, http://infolab.stanford.edu/~backrub/google.html/.18 Kolari et al. “Characterizing the splogosphere.” In Proceedings of the 3rd annual workshop on weblogging ecosystem: Aggregation, analysis and dynamics, 15th World Wide Web conference. University of Maryland, Baltimore County, 2006. http://ebiquity.umbc.edu/_file_directory_/papers/261.pdf/. As Finn Brunton writes in, Spam: a Shadow History of the Internet (Cambridge, Mass: The MIT Press, 2013), at 160, such ‘splogs’ are “a kind of PageRank greenhouse which is not in itself meant to be read by people.”19 Matt Cutts, “Another step to reward high quality sites.” Official Google Webmaster Central Blog, April 24, 2012. http://googlewebmastercentral.blogspot.ca/2012/04/another-step-to-reward-high-quality.html/.20 Chris Peterson, “Google NegativeSEO: Case Study in User-Generated Censorship.” MIT Center for Civic Media, March 20, 2013. http://civic.mit.edu/blog/petey/google-negativeseo-case-study-in-user-generated-censorship/.21 Tarleton Gillespie, “Can an Algorithm Be Wrong? Twitter Trends, the Specter of Censorship, and Our Faith in the Algorithms Around Us.” Culture Digitally. October 19, 2011. http://culturedigitally.org/2011/10/can-an-algorithm-be-wrong/.

Yet in the cases I have described, users strategically deploy their ability to arrange

information towards making specific pathways less available. The actors are amateurs, not

professionals; distributed, not centralized; they are granted power not through money or station

but through the ‘democratizing’ promise of new media. Yet rather than using this newfound

power to make information more broadly available, as the familiar story goes, they instead

deploy it to make some information harder to find, reversing their premise and their promise.

I call these cases user-generated because they are animated by signals generated by users

to overdetermine the arrangement and availability of content through their interactions with

organizational algorithms. One critical, counterintuitive insight is that the ‘social’ in ‘social

media’ is not an exclusively human sociality, but instead the ongoing and active association of

humans and nonhumans alike into a shambling mass made of people, bots, algorithms, and

other ontologically uncertain actors. I call these cases censorship following Frederick Schauer’s

ascriptive definition. Censorship, Schauer argues, is best understood as a label which “does not

describe a category of conduct, but rather attaches an operative conclusion (ascribes) to a

category created on other grounds.”22 Invocations of censorship, according to Schauer, most

fundamentally signal anxiety over the allocation of authority to determine access to or

availability of information. In adopting this definition I characterize these cases as censorship

primarily because the actors involved insistently do.

More formally, I define user-generated censorship as follows:

• Strategic interventions which suppress information by erasing or making to appear

uninteresting certain sociotechnical pathways through which it can be found

11

22 Frederick Schauer, “The Ontology of Censorship” in Censorship and Silencing: Practices of Cultural Regulation, ed. Robert Post (Los Angeles: Getty Research Institute for the History of Art and the Humanities, 1998), 160.

• Initiated neither at the behest nor the behalf of a formal public or private authority,

but instead by ‘amateurs’ empowered by the distributed mechanisms of social media

• Which, if revealed, strikes other users as an ‘unauthorized’ or ‘inappropriate’ use of

these systems, which may be ascribed as a form of censorship

If a case meets these three conditions then I believe it constitutes an emergent mode of

intervention, broadly distributed to amateurs yet empowered by centralized algorithms,

inspiring many of the same concerns and complaints of ‘traditional’ censorship but departing

from its most familiar frameworks and remedies.

3. Outline

My study of user-generated censorship proceeds in three parts.

First, I ground user-generated censorship in a case study. In Chapter II I conduct a

qualitative analysis of the Digg Patriots, a group of users active between 2009-2010 who

strategically manipulated the social news site Digg in order to make it more politically

conservative. My primary source material is an archive of nearly 13,000 posts from the listserv

through which they coordinated.23 I’ll examine and share excerpts from this corpus to develop

a better understanding of the Digg Patriots through their own words and actions as a canonical

case of user-generated censorship.

Second, I review a few models through which we might understand and evaluate the

effects of user-generated censorship. In Chapter III, I argue that user-generated censorship

could be understood a) as corrupting collective intelligence, b) as disrupting the networked

12

23 For information about the origins, controversies, and considerations of the corpus please see Appendix A.1.

public sphere, and/or c) as actor-networks assembling artifacts. The first two models, drawn

primarily from the work of James Surowiecki and Yochai Benkler respectively, are very

familiar to analyses of social media. The third, based on the work of Bruno Latour, is

somewhat less so, but I’ll discuss why I find his actor-network approach advantageous.

Third, I explore the artifacts of user-generated censorship. In Chapter IV I build on the

information science scholarship of Markus Krajewski to understand social media as

sociotechnical machines which make information available. The very centrality of these

systems, however, also makes them vulnerable points of subversion for those who would wish

to change the information made available through them. I’ll draw on some concepts from

Latour to model social media machines as producing incomplete pictures of the world which do

not reveal their own incompleteness, and to understand user-generated censorship as operating

by subtly recomposing these panoramas with persuasive effects.

Studying user-generated censorship provides a point of entry to a broader set of

questions concerning the cauldrons of conflicting expectations and uncertain effects bubbling

up from the human and nonhuman actors who together animate social media. I conclude by

discussing some of these questions and offering two paths forward for investigating the maps

made by social media.

II: A CASE STUDY IN USER-GENERATED CENSORSHIP

I don't know if someone has already said this, but Digg Patriots is like Fight Club. The first rule of Digg Patriots is you don't talk about Digg

Patriots outside of the group. The libs on digg already suspect that we are organized, but they don't know about Digg Patriots.

13

- M.D.H., “Re: BURY LIST”24

1. “Bury Until They Change Their Ways”: Introducing the Digg Patriots

Digg was founded in November 2004.25 A ‘social news’ site, Digg allowed users to

submit links to other websites and vote them up (‘digg’) or down (‘bury’). An algorithm

partially informed by these evaluations ranked the submissions in a list from most popular to

least popular. The most popular links rose to Digg’s ‘front page.’ In this way Digg’s rankings

created a feedback loop: the more popular a post became, the higher it rose, and the more

attention it attracted; the higher it rose, the more attention it attracted, and the more popular it

became. It was also a tremendous referrer of traffic to the links it ranked. By 2006, Digg was

receiving and redistributing over 30 million visitors a month, an online audience roughly

equivalent to that of the New York Times. In 2008 Digg was nearly purchased by Google for

$200 million.26 In 2010 Digg would crash, be gutted, and later reborn as a different service

under the same name, but from 2006 through 2010 it was one of the most popular and

powerful social media sites in the world. It was during this period, at the apex of Digg’s

influence, that the Digg Patriots arose.

14

24 This quote was excerpted from a text file in the Digg Patriots’ archive named 20100117-DiggPatriots-Re-BURY-LIST-10390.html. In subsequent citations I will only include the filename. 25 Lacy and Hempel. “Valley Boys.” BusinessWeek: Magazine, August 13, 2006. http://www.businessweek.com/stories/2006-08-13/valley-boys/.26 Michael Arrington, “Google In Final Negotiations To Acquire Digg For "Around $200 Million,” TechCrunch, July 22, 2008, http://techcrunch.com/2008/07/22/google-in-final-negotiations-to-acquire-digg-for-around-200-million/.

The Digg Patriots were founded in May 2009 by Lizbett and VRayZ, two highly-

active, politically-conservative Diggers. They created a Yahoo!Group, invited a small number

of trusted fellow conservatives, and began coordinating through its mailing list to advance

political conservatism on Digg, a campaign which they came to call “the cause.” R.J. Carter, an

active member and ardent supporter, described their mission as countering a perceived left-

leaning bias of Digg by coordinating their votes to swiftly and decisively bury liberal links so

that conservative stories could rise higher, making Digg more conservative in the process.

SUBJ:  Re:  Digg  Patriots

[The  Digg  Patriots  were]  comprised,  or  intended  to  be  comprised,  of  Digg  members  with  conservative  political  ideals  as  a  means  of  countering  the  left-­‐leaning  material  submitted  to  Digg  that  was  making  the  front  page  of  the  site  on  a  regular  basis....these  were  submissions  that  any  member  of  the  Digg  Patriots  would  have  marked  for  burial  upon  encounter.  But  by  organizing  under  a  Yahoo  group,  the  first  member  to  encounter  it  could  immediately  let  the  others  know  it  was  there.  This  had  the  desired  impact  -­‐-­‐  that  a  mass  of  early  burials  would  keep  it  off  the  front  page  when  the  same  number  of  burials  spread  over  time  

Figure 1: The homepage of the Digg Patriots Yahoo!Group in August 2010.

15

would  not....The  Digg  Patriots  could  have  been  more  successful  [but]  it  did  have  the  capability  of  keeping  the  field  clear  of  "debris  and  detritus"  so  that  other  news  stories  could  get  to  the  front  page.  

 -­‐  R.J  Carter,  February  12  201327

SUBJ:  Re:  DiggPatriots-­‐  Whining  againI'll  continue  to  bury  their  submissions  until  they  change  their  ways  and  become  conservatives.  =^)

-­‐  VRayZ,  June  6  201028

At least 54 members of the Digg Patriots exchanged almost 13,000 messages over the

course of 11 months. The corpus includes almost 7,000 unique Digg links they targeted,

although this represents a tiny fraction of their total activity. From April 2009 to August 2010

Lizbett alone dugg an astounding 76,000 links and submitted over 1,600 articles, 20% of which

made the front page, a respectable ratio for a site with millions of members and billions of

submissions.29 Many of their campaigns, like this exchange from early in the corpus, began

with a link to be targeted, with subsequent responses confirming action had been taken.

SUBJ:  BuryKill  stab  hit  spit  bury  thank  youhttp://digg.com/political_opinion/Dickipedia_Teabaggers_2

-­‐  Lizbett,  September  22  200930

SUBJ:  Re:  Burystomped  on  it  with  golf  cleats...

-­‐  Individually  Warped,  September  22  200931

16

27 R.J. Carter, email to author, February 12, 2013. As I describe in Appendix A.1, R.J. was one of only two Patriots to respond to my requests for comment. This quote was in response to an email from me saying “I'd be curious to know how you and others defined the movement, its validity, and its success (or failure).”28 20100606-DiggPatriots-Whining-again-1877.html.29 For Lizbett’s statistics, see http://socialblade.com/digg/top1000users.html under name ‘bettverboten’. Statistics for her original account, ‘lizbett,‘ can be seen at http://socialblade.com/digg/topgraveyard.html. Both links accessed on May 24, 2013. 30 20090922-DiggPatriots-Bury-8405.html31 20090922-DiggPatriots-Bury-8404.html

SUBJ:  Re:  Burystomped  flat

-­‐  Gloria  L,  September  22  200932

SUBJ:  Re:  BuryNot  only  that,  I  reported  it  as  spam,  pornographic,  and  "calculated  hate  that  goes  beyond  the  limits  of  taste  even  in  political  invective."

I  also  pointed  out,  as  if  they  didn't  know,  that  "jblogging"  and  "thejoshuablog"  are  one  and  the  same  accountholder,  and  cited  his  avatar  as  evidence.

-­‐  Temlakos,  September  22  200933

By far the most common action was downvoting, but, as can be seen in Temlakos’ post,

the Patriots would also report liberal posts as spam, duplicates, or for other Terms of Service

violations. In addition to sending messages to each other, they also employed a variety of tools

and techniques which allowed them to closely track submissions by liberal users so that they

could instantly target them for downvoting.

Figure 2: An message from Lizbett targeting an link (ironically about increasing conservative influence on Digg) for burying.

17

32 20090922-DiggPatriots-Bury-8403.html33 20090922-DiggPatriots-Bury-8402.html

2. “Any Chair In A Bar Fight”: Mission, Metrics, & MethodsSUBJ:  Going  to  be  very  busy  for  the  rest  of  the  month...Keep  beating  back  the  flood  of  stupidity,  also  known  as  liberalism.  Fight  the  good  fight.  

-­‐  tccbossm4n,  July  16  201034

SUBJ:  Re:  Renewed  Interest  in  BuryingI'm  amazed  at  their  lack  of  organization.    If  they  had  half  a  wit  among  them  they  could  bury  any  and  every  submission  we  have  but  as  in  real  life  I  guess,  liberals  just  roll  with  the  tide  through  Digg  like  they  do  in  real  life.    They  have  no  work  ethic,  no  core  values  and  no  common  sense  beneath  the  paper  thin  liberal  skin  they  cover  themselves  with.      

-­‐  VRayZ,  June  17  201035

  “I’d love for you to feel better about Digg and a be stronger player” Lizbett encouraged

one new member.36 Indeed, many Patriots focused on becoming ‘stronger,’ ‘more effective’

Diggers, which signifies how profoundly they thought of Digg as a system through which they

could do something. They needed to develop strength, to become stronger, to win the battles in

Figure 3: A screenshot circulated among the Patriots celebrating they had successfully buried four of the five most recent articles submitted by a liberal user they tracked.

18

34 20100716-DiggPatriots-Going-to-be-very-busy-for-the-rest-of-the-month-not-around-much-5884.html35 20100617-DiggPatriots-Renewed-Interest-in-Burying-1786.html36 20100621-DiggPatriots-What-should-I-do-2421.html

which they believed they were embroiled. “We need to continue the fight,” read the first line of

the Yahoo!Group’s description, and such combative language pervades the Patriots’ posts.

“Another one bites the dust,” gloated tccbossm4n of a post they successfully buried.37 The

Patriots believed that they were outnumbered by liberal users but could outwork and outfight

them to make Digg more conservative.

SUBJ:  Re:  Bury  Now  Novahater's  subAgain  the  question  arises  about  the  validity  of  us  organizing  through  email...

I  feel  we  are  far  outnumbered.  So  does  that  make  what  we  do  right?To  fight  for  what  is  right  and  just,  I  would  say  yes.  Hopefully  more  people  will  see  our  beliefs  as  the  right  way.  We're  called  the  right  for  a  reason.

-­‐  R.J.C.,  October  29  200938

SUBJ:  Re:  Bury  Now  Novahater's  subI  think  Digg's  rules  against  networking  to  organize  digging  and  burying  are  bulls**t,  so  I  don't  have  any  remorse  about  our  own  networking.    As  for  trying  to  get  the  other  side  in  trouble  for  doing  it-­‐-­‐I'm  a  big  believer  in  the  "any  chair  in  a  bar  fight"  philosophy.

Kurt  H.,  October  29  200939

Digg’s Terms of Service forbade “inflating or altering the ‘digg count’ [or] participating

in any other organized effort” to affect rankings, and the Patriots themselves believed they

were breaking these rules through their coordination. They were very worried they would be

caught: “Be very careful who you invite here, as the ‘bury’ list can get ALL of us permanently

19

37 20090927-DiggPatriots-Re-Another-Noupsell-Bury-Please-8525.html38 20091029-DiggPatriots-Bury-Now-Novahaters-sub-14422.html39 20091029-DiggPatriots-Bury-Now-Novahaters-sub-14423.html

banned,” warned Janet M. However, they justified their rulebreaking by appealing to the

conservative cause - and their love of fighting liberals.

SUBJ:  Re:  let’s  see  if  he  takes  the  bait...I'm  a  big  mouth.    I  did  not  join  Digg  (and  I  did  long  before  I  came  here)  to  be  a  shrinking  violet.    I  didn't  join  in  order  to  put  submissions  on  the  front  page.

I  joined  in  order  to  be  myself  and  to  give  some  a-­‐hole  liberals  who  had  been  unchallenged  ...  I  joined  to  give  them  a  challenge.

-­‐  Jill  K.,  June  21  201040

SUBJ:  Re:  DiggPatriots-­‐  Maybe...if  digg  loses  it's  competitive  nature,  and  let's  face  it  the  real  satisfaction  [is]  in  burying  the  fools  and  hearing  them  cry  endlessly  about  it,  where  is  the  fun  ?  the  whole  "everyone  wins  because  the  only  people  i  will  relate  with  agree  with  me"  thing  is,  how  can  i  say  it,  too  freakin'  libtard  for  me.  i  don't  use  my  twitter,  facebook,  or  myspace  accounts  and  landed  on  digg  because  i  like  fighting  with  my  enemies  and  i  really  like  winning.  The  dp's  have  had  an  impact.  where  will  the  impact  be  if  you're  swimming  with  the  current  ?

Brendan  M.,  July  5  201041

“Half the reason I come to digg is to bury the libs,” observed VRayZ.42 Indeed, a love of

fighting for its own sake made it worthwhile for many Patriots to battle even the most

incorrigibly liberal opponents. “[They] are what they are and they aren't going to change. bury

them without remorse,” wrote Brendan M. However, even he still evaluated his actions by

whether they advanced the cause - whether they had an ‘impact.’43 Brendan M., like many

other members of the Patriots, stayed at Digg in part because he thought it was too liberal and

because he thought he could make it more conservative. He lamented the rumored removal of

20

40 20100621-DiggPatriots-lets-see-if-he-takes-the-bait-2591.html41 20100705-DiggPatriots-Maybe-4350.html42 20100628-DiggPatriots-anamolys-mad-at-me-3447.html43 20100617-DiggPatriots-Whining-again-1911.html

the bury button, which would remove his ability to have an impact, to make a difference. When

the love of fighting ran up against the need to win, it was usually the latter which carried the

day.

SUBJ:  Re  -­‐  BurY  Mutual?Please,  Please,stop  the  discussions.    You  are  playing  right  into  their  hands.If  you  just  can't  help  yourself,  then  Maybe  you  should  find  another  outlet  for  your  frustration.  I  spend  far  too  much  time  on  Digg  to  see  it  wasted  by  immature  sniping.I  hope  no  one  is  offended,  but  remember  why  we  are  here.  We  want  to  Depress  the  progressive  stories,  while  encouraging  conservative  ones.

-­‐  rgcmsg  G,  March  29  201044

The dynamic tension between these contradictory motivations gave rise to a fascinating

rule against commenting on liberal posts. “Don't comment while it's live (wait 24hr or close

with not many diggs) is the general rule of thumb,” advised asamidigg.45 This convention

developed after the Patriots came to believe the Digg algorithm interpreted comments as an

index of interest: posts with more comments would rise more quickly, and posts with less

would fall further faster. They treated comments not as a process of deliberative discourse but

as a tool which could be leveraged to increase or decrease visibility. Instead of raging at their

enemies, they did their best to keep quiet: “Venting here, so I don't violate DP discipline and

comment on a submission we're trying to bury,” as Kurt H. wrote.46 Meanwhile, they tried to

trick liberal users into pushing their own submissions higher. “We probably should concentrate

on commenting on each others stories and creating conversational comments to pull in more

people to comment when they come by to digg on our stories,” mused Lizbett.47 “I try to do it

21

44 20100329-DiggPatriots-BurY-Mutual-12404.html45 20100325-DiggPatriots-Re-Bury-12212.html46 20100326-DiggPatriots-Bury-12308.html47 20100212-DiggPatriots-Re-Novahators-Sub-More-BURY-13994.html

for all the DPs subs, especially if there are no comments, I comment just to try and start some

conversation.”48

SUBJ:  Maybe  we  can  still  bury  thiswhat's  with  [liberal  Digger]  PhilPerspective  hitting  the  front  page  so  much  lately?  Looks  like  we're  falling  down  on  the  job,  or  at  least  need  to  shift  our  focus.  

-­‐  tccbossm4n,  December  4  200949

SUBJ:  Anomaly100  thinks  Novenator  is  all  that...What  she  doesn't  know,  is  that  our  group  is  growing  larger  by  the  day  and  [liberal  Digger]  novy  is  going  to  have  a  harder  time  having  his  stories  pop...

Individually  Warped,  February  14  201050

  The Patriots evaluated their impact by straightforward metrics: they wanted

conservative posts to “pop” (rise to the front page) and liberal posts to be “buried” (sunk deep

in the rankings). Popped stories were alive with attention; buried stories were effectively dead.

“We are the Kevorkian of liberal stories on Digg. I hate Kevorkian, but I think in this case it is

a great comparison. It is like putting down a suffering animal,” joked Individually Warped.51

They bent their collect effort to make it harder for liberal links ever to rise from the dead.

SUBJ:  RE:  BuryYou  know  people....    I  got  a  feeling  if  we  can  pull  another  10  or  15  active  DP  members  in  we  could  turn  Digg  conservative.....

VRayZ,  March  5  201052

SUBJ:  Re:  I  think  we  need  to  bury  thisIt  is  kind  of  a  learn  as  you  go  deal.  I    see  what  works  for  other  people  so  I  try  to  reshape  my  actions  to  make  me  a  better  social  digger.  

22

48 20100413-DiggPatriots-I-think-we-need-to-bury-this-12731.html49 20091204-DiggPatriots-Maybe-we-can-still-bury-this-9649.html50 20100214-DiggPatriots-Anomaly100-thinks-Novenator-is-all-that-11285.html51 20100628-DiggPatriots-anamolys-mad-at-me-3441.html52 20100305-DiggPatriots-Bury-11765.html

Lizbett,  April  13  201053

But the Patriots knew they weren’t strong enough to win this fight on their own. “DPs

is a place to share and get support for your subs, but we are not big enough or strong enough

to get the job done,” observed Lizbett.54 However, they were extremely hesitant to bring new

members on board who might betray their coordination and get them banned. “[He] could be

like a digg double agent, wanting us to think he's on our side when all along he's playing for

the other team,” worried allisonrose870 of one proposed recruit.55 Over the course of the

corpus only two new members were added to the group, and only after having been thoroughly

and carefully vetted by senior members. To become stronger players, as Lizbett urged all

Patriots to do, they instead invented a way to enlist reliable, riskless recruits to join their team.

SUBJ:  RE:  FP  for  J!Okay  folks,  want  to  hit  FP  often?  Follow  J's  lead.  He's  got  90+  friends  all  who  digg  early  (this  is  key).  If  you  can  cultivate  90-­‐100  friends  like  this  your  subs  will  hit  on  a  regular  basis,  but  cultivating  this  many  GOOD  friends  requires  you  to  do  the  same  for  them.  Gotta  digg  'em  early  and  never  miss.

phil  d.,  June  17  201056  

SUBJ:  Re:  DiggPatriots-­‐  What  should  I  do?For  the  next  2/3  weeks,  don't  sub  at  all,  but  concentrate  on  a)  digging  good  mutuals(ie  those  who  digg  more  than  50%  of  your  stuff)  AND  also  b)  concentrate  on  digging  active  diggers...  It's  important  to  remember  that  digg  is  all  about  give  and  take.    If  you  never  digg  someones  subs,  they  will  notice  and  not  bother  digging  yours.  Conversely,  if  they  notice  you  digging  theirs,  they  will  digg  yours...

23

53 20100413-DiggPatriots-I-think-we-need-to-bury-this-12731.html54 20100621-DiggPatriots-What-should-I-do-2421.html55 20100712-DiggPatriots-Concerning-Dilberto-5276.html56 20100617-DiggPatriots-FP-for-J-1836.html

 Change  your  profile  to  mention  that  are  an  active  digger  and  you  faithfully  digg  your  mutual's  subs.  Keep  your  mutuals  list  to  somewhere  between  90  and  130  mutuals.  This  will  enable  you  to  get  all  your  mutual's  subs  early  and  reliably.  

Remember  that  digg  is  all  about:  'you  scratch  my  back  and  i'll  scratch  yours'.

It  may  take  2  or  3  months  for  you  to  get  yourself  into  the  consciousness  of  active  diggers  around  digg.  For  each  FP  that  you  get,  that  increases  your  visibility  and  the  'marketing'(when  you  comment  early  on  people's  subs  and  you  digg  other  people's  subs  early,  you  are  effectively  marketing  yourself)  of  yourself  many-­‐fold.

Get  rid  of  those  in  your  mutuals  list  who  are  below  your  threshold  (eg  less  than  4  or  5  of  your  recent  subs)  and  mutuals  those  in  your  fans  list  who  are  digging  your  stuff  often  and  early.

-­‐  J.,  June  21  201057

  The solution was hit upon by J., a sort of senior strategist and tactician for the Patriots

whom Lizbett fondly described as her mentor. In 2009-2010 Digg provided a symmetrical

relationship between users called “mutuals”: if two users agreed to be mutuals then each could

easily see what the other had submitted. Digg’s developers designed mutuals to assist users in

seeing what their friends were sharing as a means of facilitating serendipitous discovery. J.,

however, realized the Patriots could extend their influence through mutuals - if they picked the

right ones. He began grooming mutuals, selecting them not by whether they agreed with his

politics, but whether they would rapidly, regularly, and reliably upvote his submissions.58 By

doing so he was able to muster an army of supporters to push his stories higher. It proved an

24

57 20100621-DiggPatriots-What-should-I-do-2439.html58 Compare the Patriots’ use of mutuals to Brunton’s observations on the phenomenon of ‘personality spamming,’ in Spam, 165: “the work of arrogating attention for oneself, using social media to build an audience -- often a very carefully quantified audience of ‘followers’ and ‘rebloggers’ -- rather than a network of friends, as was the initial, notional promise. It is a witty condemnation of the socially acceptable but aggressively eyeball-hungry work of those who would be, or act like, celebrities, ‘influencers’ or ‘thought leaders.’”

extraordinarily successful strategy and other Patriots quickly followed suit. “It took me over a

month of cultivating mutuals and friends who digg daily to build up to 125+ diggs a sub,”

observed phil d., who cheerfully noted that he was now recruiting hundreds more.59

SUBJ:  BURY...I'm  sure  it's  the  same  with  most  people  as  it  is  with  me  -­‐  things  are  rarely  personal  on  digg.  The  only  thing  that  matters  is  if  you  are  digging  their  stuff  or  not....There  is  a  feminist  envirnmentalist  in  my  list(tomboys),  but  she  seems  to  be  quite  good  at  digging  my  stuff  so  i  don't  care  one  iota  what  her  beliefs  are.

-­‐  J.,  October  30  200960

SUBJ:  Re:  Bury...I'm  the  same,  Alan.  The  only  thing  i  care  about  is  if  they  are  going  to  digg  my  stuff  consistently  early  or  not.  I  don't  care  if  they  are  a  rabid  environmentalist  or  a  6  humped  creature  from  planet  Zog.  If  they  digg  my  stuff  and  are  ultra  active,  then  they're  in  my  list.

-­‐  J.,  December  19  200961

SUBJ:  Re:  Please  help  with  comment  diggs  Thank  youTo  have  any  chance  to  get  some  of  these  articles  to  FP,  we  are  going  to  have  to  push  extra  hard...I've  kept  a  low  profile  (politically)  to  build  up  a  strong  mutual  base,  which,  as  you  know,  we  need.  NOW  is  the  time  to  see  who  will  digg  us,  after  months  of  digging  all  their  TRASH  (I'm  sure  you  know  what  I  mean)  

Janet  M.,  March  22  201062

There was broad agreement among the Patriots that mutuals should be measured by

how effectively they extended conservative influence. Even Janet M., an ultraconservative

FreeRepublic poster who bitterly hated liberals, assumed a ruthlessly pragmatic perspective. “I

don't befriend based on someone's politics. I look at their stats,” wrote alanocu_digg.63 J.

25

59 20100606-DiggPatriots-Looks-like-we-missed-one-8307.html60 20091029-DiggPatriots-Bury-8916.html61 20091219-DiggPatriots-Re-Bury-9902.html62 20100322-DiggPatriots-Please-help-with-comment-diggs-Thank-you-12118.html63 20091219-DiggPatriots-Re-Bury-9896.html

advised the Patriots not to mutual anyone “east of the Middle East,” not out of latent

Occidentalism but because the time difference would negatively impact their ability to rapidly

upvote his submissions.64 For the Patriots, the most valuable mutuals resembled a kind of

mathematical function or mechanical device, a black box into which new submissions were fed

and out of which more attention emerged. In a striking ontological inversion, they evaluated

human mutuals by their least human characteristics; as tools which were useful as long as they

‘worked’ and discarded as broken once they did not. And at the same time as the Digg Patriots

began to evaluate humans as technology, they began to evaluate technologies as human.

SUBJ:  Who  Runs  Digg?...There's  no  'algorithm'  at  digg.  The  'algorithm'  most  likely  consists  of  a  bunch  of  liberal,  bi-­‐sexual,  emo-­‐types,  who  drink  mimosas  all  day,  and  engage  in  a  circle-­‐jerk  by  night.  When  they're  not  doing  that,  they  pull  a  few  levers  to  get  a  banana  payoff  from  a  machine,  which  they  call  the  digg  'algorithm'.  Just  my  opinion,  but  prolly  spot-­‐on...

-­‐  rjwusa,  July  20  201065

SUBJ:  Re:  Digg’s  Reply  -­‐  They  buried  my  Sub  in  2hrs...It  IS  a  bunch  of  shit.  My  sub  today  did  not  pop  with  300+  diggs  and  45  comments  So  I  wrote  them  and  asked  them  why  are  you    ignoring  my  stories  (I  have  done  this  at  least  10  times  now  so  it  is  not  like  it    is  the  first  time)  Well  low  and  behold  my  story  popped  about    10  minutes  later  about  an  hour++after  the  24  hour  periodIf  I  do  nothing  they  do  nothing.  If  I    complain  my  story  pops.  If  they  like  you...  your  stories  pop  with  100  or  less  diggs  and  maybe  4  comments.  I  like  playing  the  digg  game  but  my  faith    in  any  REAL  algorithm  is  gone.Some  people  like  to  always  believe  in  Santa  Claus  too,  but  I  like  common  sense  and  reality...

-­‐  Lizbett,  March  25  201066

26

64 20091029-DiggPatriots-Bury-8921.html65 20100719-DiggPatriots-Who-Runs-Digg-6336.html66 20100326-DiggPatriots-Diggs-Reply-They-Buried-my-Sub-in-2hrs-12276.html

SUBJ:  Re  -­‐  Bury  if  you  haven’t.  It’s  FPAnd  as  I  predicted  earlier:"This  [liberal]  sub  just  past  the  10  minute  mark...  It  will  probably  stay  down  for  about  another  5  to  10  hours  and  then  the  A.M.A.(Algorithm  My  Ass)  will  kick  in  and  promote  this  story...

-­‐  Individually  Warped,  March  26  201067

  The opaque Digg algorithm was a source of constant fascination and frustration for the

Patriots. When liberal posts they thought safely buried popped back to the top, or conservative

posts they backed with their collective forces failed to rise, the Patriots suspected that the Digg

algorithm was at best broken, and at worst a sham, a technological facade which laundered the

San Franciscan sympathies of its engineers. Their complaints could be read as admittedly

vulgar versions of Winnerian critiques of the politics of artifacts: the unpredictable output of

algorithms anthropomorphized into human capriciousness, as liberal politics embedded in the

infrastructure of code.68 Some of them came to believe that the algorithm was even a lie. “I

think there are no ‘readers’ claiming that it's inaccurate, but digg admin themselves. I also

think that it won't reach front page because digg admin won't let it,” suspected J.69 This is an

especially striking observation because it was precisely J. who, in other situations, so clearly

envisioned Digg as a tool he could reliably repurpose. Yet even for him there was no single,

essential algorithm, but rather many algorithms imagined differently depending on whether it

was behaving as they expected.

27

67 20100326-DiggPatriots-Bury-if-you-havent-Its-FP-12302.html68 Langdon Winner, “Do artifacts have politics?” Daedalus 109, no. 1 (1980): 121-136.69 20091002-DiggPatriots-Its-not-just-conservatives-who-get-banned-8533.html

3. “A Rude Wake-Up Call”: Reading the Digg PatriotsSUBJ:  Re:  Digg  PatriotsBeing  a  popular  Digg  user  was  fun.  But  it  was  a  lot  of  work  too.  I  am  not  sure  of  what  my  rank  was  before  the  site  was  sold  and  changed  formats,  but  I  was  in  the  top  200.  There  were  others  in  the  group  with  higher  rankings  than  mine.  To  get  a  story  popular  on  Digg,  you  had  to  follow  other  Digg  power  users  loyally  and  upvote  or  "digg"  their  submissions.  The  sweet  spot  was  about  200  to  300  mutual  friends.  The  reason  I  am  telling  you  this,  is  because  the  most  successful  members  of  the  DiggPatriots  were  friends  with  all  types  of  people  who  were  just  as  dedicated,  just  as  addicted,  just  as  loyal  to  the  site  but  were  in  no  way  conservative.

Getting  a  submission  to  the  front  page  of  Digg  meant  that  hundreds  of  thousands  of  people  would  see  your  submission  and  comment  on  it.  Once  I  started  to  get  the  knack  of  the  site,  I  could  get  a  front  pager  nearly  every  day.  

What  was  the  best  thing  to  come  out  of  Digg?  I  became  really  close  friends  with  all  the  active  members  of  DiggPatriots.  I  never  met  them  in  real  life,  but  I  talk  

Figure 4: A screenshot of the Patriots’ Yahoo!Group file directory. Note the guides and tools posted by chroniccolonic.

28

with  them  on  the  phone  and  I  consider  them  to  be  very  good  friends.  We  are  still  really  good  friends  to  this  day.If  it  were  not  Digg,  we  would  not  have  known  each  other.  For  that,  I  could  not  thank  the  creators  of  Digg  enough.

-­‐  ChronicColonic,  February  23  201370

SUBJ:  Re:  Digg  PatriotsIn  short,  the  Digg  Patriots  was  a  response  to  a  wave  of  leftism  that  had  grown  smug  and  arrogant  in  its  presence  on  Digg.  We  were  a  rude  wake-­‐up  call  that  there  were  those  who  felt  otherwise  within  Digg.

-­‐  R.J.  Carter,  February  12  201371

By August 2010 the Patriots were planning to leave Digg. A proposed redesign, created

in part to help stop what administrators called “bad behavior like group-burying,” eliminated

the bury button and mutuals, thwarting the Patriots’ methods. The group began to shift to

other venues, and on August 1st, 2010, phil d. targeted their first link on reddit.72 Later that

week a muckraking blogger named OleOleOlson - who was also the Patriots’ most hated foe

on Digg - published a post entitled “Massive Censorship of Digg Uncovered.”73 He accused

them of “a widespread campaign of censorship” and published excerpts from their listserv to

prove it.74 The story drove the Patriots deep underground; their archive breaks off as sharply

as it began, and there is no reliable record of their continued activity.

What do we know of the Digg Patriots? They thought they could make Digg and, by

extension, other users more conservative by gathering together not only their own members

29

70 ChronicColonic, in Facebook message to the author. 71 R.J. Carter, in email to the author. 72 20100801-DiggPatriots-If-youre-on-Reddit-7971.html73 OleOleOlson, “Massive Censorship of Digg Uncovered,” Alternet, August 5, 2010, http://blogs.alternet.org/oleoleolson/2010/08/05/massive-censorship-of-digg-uncovered/.74 Ironically, Olson himself was an unabashed Digg strategist, who proudly published guides for how progressives could use mutuals to extend their message just as the Patriots had. See, e.g., “Progressives Guide to Social Media 3: Digg,” http://newsjunkiepost.com/2012/02/14/progressives-guide-to-social-media-3-digg/. The few Digg Patriots I was able to reach were all angry that Olson, whom they believed manipulated Digg just as much as they, had not only been able to successfully cast them as censors, but had escaped similar blame him.

but also mutuals and the affordances of the algorithm. They were profoundly social, but not

only with humans, but with technologies, and humans which they treated like technologies:

they battled against humans (liberals) and algorithms (as liberal) with the assistance of humans

(as tools) and algorithms (as tools). Their exact impact is uncertain, but there is no doubt that

they made Digg different by fighting to make liberal content less visible than it would have been

without their intervention.75

This complicated legacy could be read in several different lights. Their urge to battle

liberals as underdogs could be placed historically in the context of the decline of a particular

brand of conservatism. Their understanding of Digg as a manipulable system could be situated

culturally in the context of comparative privilege providing a sense of strategic agency.76 Their

need to cooperate to achieve their goals could be considered paradoxically in the context of

their strongly individualistic ideology. Their efforts to suppress and silence, rather than

promote and engage with, expression could be read ironically in the context of their frequent

‘patriotic’ references to the Constitution.

In this thesis, however, I’m interested in understanding how Olson, along with

countless others, could so easily read the Patriots as censors, for this interpretation tells us less

about the Patriots proper than it does about the interpreters themselves. Why were the Patriots

almost universally understood as bad actors subverting Digg and not (say) as an oppressed

minority resisting hegemony? What makes group burying ‘bad behavior’? Why did Digg

30

75 The impact of the Patriots is difficult to measure for several reasons. First, because of the many redesigns Digg has undergone, I have been unable to find any data from when they were active. Second, it’s not clear how their impact would be measured even if such data could be found. What would be the baseline? I elaborate upon this question further in Chapter III in the section on collective intelligence. 76 Critical legal scholars often observe that different individuals have different perspectives on their relationship with law depending on their situation within society. See, e.g., Silbey & Ewick, The Common Place of Law: Stories from Everyday Life (Cengage Learning, 2005), particularly chapter 5.

forbid it? If the Patriots were ‘breaking’ Digg, what exactly were they breaking, and why?

Where, whom, or from what do these expectations arise that the Patriots could violate them?

III: UNDERSTANDING USER-GENERATED CENSORSHIP

The interesting question is not whether Twitter is censoring its Trends list. The interesting question is, what do we think the Trends list is, what it

represents and how it works, that we can presume to hold it accountable when we think it is “wrong?” What are these algorithms, and what do we

want them to be?- Tarleton Gillespie, “Can an Algorithm Be Wrong?”77

1. As Corrupting Collective Intelligence

One way to model the impact of the Patriots is as corrupting the collective intelligence

which would otherwise be elicited by Digg. Social media have long been popularly understood

as systems for aggregating the wisdom of many minds: “an essential part of Web 2.0,” wrote

Tim O’Reilly in 2005, “is harnessing collective intelligence, turning the web into a kind of

global brain...”78 Platforms like blogs and social media, he wrote, lowered the barriers to

publish to the Web, which permitted more people to post and link between content with

powerful emergent effects. According to O’Reilly, “much as [Google’s search algorithm]

PageRank produces better results than analysis of any individual document, the collective

attention of the blogosphere selects for value.”79

Perhaps the most popular contemporary writer on collective intelligence is James

Surowiecki. In his book The Wisdom of Crowds Surowiecki argues that, under certain conditions,

markets, or systems which behave like markets, can elicit some aggregate collective intelligence

31

77 Gillespie, “Can An Algorithm Be Wrong?”78 Tim O’Reilly, “What Is Web 2.0?” O’Reilly Media, September 30, 2005, http://oreilly.com/web2/archive/what-is-web-20.html?page=3/.79 O’Reilly, “What Is Web 2.0?”

superior to that distributed among its constituent members.80 Surowiecki himself believed new

media to offer opportunities for the ‘wisdom of crowds’ to emerge: collective intelligence, he

argued, is “the reason the Internet search engine Google can scan a billion Web pages and find

the one page that has the exact piece of information you were looking for.”81

But not all crowds are wise, and not everything collected by them constitutes

intelligence. Surowiecki identified “four conditions that characterize wise crowds:”82

• Diversity of opinion: each person should have some private information, even if it’s

just an eccentric interpretation of the known facts

• Independence: people’s opinions are not determined by the opinions of those around

them

• Decentralization: people are able to specialize and draw on local knowledge

• Aggregation: some mechanism exists for turning private judgments into a collective

decision

Surowiecki’s pop scholarship has distinguished intellectual ancestry. He traces his work

directly to Frances Galton, but another forebear could well have been the Marquis de

Condorcet, who advanced a statistical argument for democracy in 1785. His eponymous jury

theorem in “Essay on the Application of Analysis to the Probability of Majority Decisions”

proved, as a matter of math, that if every voter in a given pool is more likely than not to vote

‘correctly,’ the probability that the majority vote is ‘correct’ increases with the size of the pool

32

80 James Surowiecki, The Wisdom of Crowds (Random House Digital, Inc., 2005). 81 Surowiecki, Wisdom of Crowds, 15. 82 Surowiecki, Wisdom of Crowds, 10.

(and decreases if voters are more likely to be ‘wrong’).83 Surowiecki’s contemporaries include

Cass Sunstein, who adopted the Condorcet theorem, along with Hayek’s price theory and

other tools from economics and political science, to divine the conditions under which certain

systems, particularly ‘Web 2.0’ systems like Slashdot and Wikipedia, may elicit ‘correct’

answers.84

These models of collective intelligence share some common premises. First, for at least

some questions a ‘correct’ answer exists, and at least part of this knowledge is distributed

among individuals. Second, this distributed knowledge, which Surowiecki calls “private

information,” can be aggregated if many individuals are allowed to make decisions

independently. Independence is a key criterion because it allows crowds to solicit private

information from many uncoordinated sources such that their individual ‘errors’ do not

correlate and compound.85 The more private information the better, since, in theory at least,

33

83 See “Condorcet’s jury theorem,” Wikipedia: The Free Encyclopedia, accessed May 24, 2013, http://en.wikipedia.org/wiki/Condorcet's_jury_theorem/.84 Cass Sunstein, Infotopia: How Many Minds Produce Knowledge (Oxford University Press, 2006). 85 Of course, a premise to this premise is that sufficient autonomy is even possible: Surowiecki, who calls independence “intuitively appealing” and “at the core of Western liberalism,” admits it is “hard to come by” in everyday life, where social scientists tend to “describe people as embedded in particular social contexts, and see influence as inescapable.” Surowiecki, Wisdom of Crowds, 42.

inaccuracies should cancel out such that only ‘good’ information remains, like salt left behind

by evaporated seawater.86

For Surowiecki and Sunstein, collective intelligence is disrupted by foreknowledge or

coordination, either of which corrupts private information and produces information cascades.

Information cascades, Surowiecki writes, occur “when people’s decisions are not made all at

once but rather in sequence, so that some people go to one of the [options] first and then

everyone else follows in order,” transforming decisions into “a sequence of uninformed choices,

so that collectively the group ends up making a bad decision.”87 When actors don’t act

independently, or when there is insufficient diversity of opinion, collective intelligence is

impaired; as with Condorcet, it is reliably unreliable.

From this perspective the Digg Patriots can be seen to corrupt the conditions necessary

for collective intelligence to reliably emerge. Surowiecki’s model requires humans to behave

with naive, autonomous independence; indeed, he specifically acknowledges ‘situatedness’ as a

problem for collective intelligence. The Patriots and their mutuals, however, constituted a

coordinated conservative bloc which colluded to form information cascades. Kristina Lerman

34

86 Another interesting tradition in eliciting ‘correct’ information emerged, not from the humanities and social sciences, but from electrical engineering. In his 1937 masters thesis for MIT, which has been called the most importance of the century, Claude Shannon demonstrated that circuits could be built to evaluate and solve boolean logic; Shannon, “A symbolic analysis of relay and switching circuits,” (S.M. thesis, Massachusetts Institute of Technology, 1937), http://dspace.mit.edu/handle/1721.1/11173. In 1949 he copublished The Mathematical Theory of Communication and coined the term “bits” to designate the binary digits which could be used to express information through their boolean state; Shannon & Weaver, The Mathematical Theory of Communication (University of Illinois Press, 1949). Shannon, a researcher at Bell Labs, was primarily trying to figure out how to make telephone communications more efficient, but James Gleick argues that his work represented a turning point in the history of information; Gleick, The Information: A History, A Theory, A Flood (Vintage, 2012). It was Shannon, Gleick argues, who provided the theoretical base for the idea that ‘communication’ could be understood most fundamentally as the transmission of information. Under this model, Gleick observes, “information is surprise,” is what is new; see Gleick, The Information, 284. The ‘signal’ is what is left behind when the ‘noise’ has been canceled out through an algorithmic process, which for Shannon was math, but could also be seen as voting on Digg or other social media. The parallels, especially to Surowiecki’s arguments about public vs private information, are striking, but the discourses different enough that I have not developed the argument further in this paper.87 Surowiecki, Wisdom of Crowds, 55.

has demonstrated that “social networks [within Digg] have a strong impact on the number of

votes received,” and the Patriots maintained armies of mutuals to achieve precisely this

dispositive effect.88

But while seeing the Patriots as disrupting popular models of collective intelligence is

straightforward, it’s also not necessarily the most productive perspective to adopt. It’s been

well-documented that social media aren’t particularly effective aggregators of collective

intelligence under the best of circumstances, as Surowiecki himself pointed out when reddit

persistently misidentified suspects after the bombing of the Boston Marathon.89 In a striking

study of over 170,000 links posted to reddit, Eric Gilbert found that over 52% of the links

which became popular were in fact resubmissions which had failed to initially become popular,

demonstrating that the community was a reliably unreliable arbiter of its own interests.90

There are two other reasons that the collective intelligence model, while still very

popular in the social media context, is not a particularly helpful guide to understanding user-

generated censorship. First, it presumes that knowledge exists ‘out there’ in the world to be

elicited by processes, which runs against the proposition that knowledge is instead constructed

through processes. It’s not a priori obvious that the Patriots should be evaluated by to what

degree they made Digg more or less ‘correct’ as much as that they made Digg different. Second,

because the collective intelligence model eschews coordination in favor of independence, it

doesn’t provide a conceptual apparatus by which to understand the Patriots as a strategic

35

88 Kristina Lerman, “Social information processing in news aggregation,” Internet Computing, IEEE 11, no. 6 (2007): 16-28; Lerman & Galstyan, “Analysis of social voting patterns on digg.” In Proceedings of the first workshop on Online social networks, pp. 7-12. ACM, 2008, at 8. 89 James Surowiecki, “The Wise Way to Crowdsource a Manhunt,” The New Yorker: Daily Comment, April 24, 2013, http://www.newyorker.com/online/blogs/comment/2013/04/reddit-tsarnaev-marathon-bombers-wisdom-of-crowds.html/.90 Eric Gilbert, “Widespread underprovision on Reddit.” In Proceedings of the 2013 conference on Computer supported cooperative work, pp. 803-808. ACM, 2013, http://comp.social.gatech.edu/papers/cscw13.reddit.gilbert.pdf/.

group; it can only characterize them as ‘bad actors.’ These are not, of course, problems with

collective intelligence as a model, but they may be indications that collective intelligence might

not be the most useful way to think about user-generated censorship.

2. As Reintroducing Heteronomy to the Networked Public Sphere

Another way to model the Patriots is as reintroducing heteronomy to the networked

public sphere. The concept of the networked public sphere was introduced by Yochai Benkler

in Wealth of Networks as a component of his contention that the Internet constitutes a

revolutionary advance for human freedom.91 Before discussing the Patriots’ effect on the

networked public sphere, however, it may be helpful to explore Benkler’s ideas a bit. Though

the concept of the public sphere is most strongly tied to the work of Jurgen Habermas, whom

Benkler frequently cites, the line tying the two is actually deceptively long and twisty, as

Benkler’s theory of the public sphere departs from Habermas’ in several important respects.

First, for Habermas, a “public sphere from which specific groups would be eo ipso

excluded was less than merely incomplete; it was not a public sphere at all.”92 This necessary

condition of “universal access” was most famously critiqued by Nancy Fraser when she

hammered Habermas for allegedly ignoring the exclusion of women from his historical

examples of the public sphere.93 Benkler, however, readily acknowledges that the Internet is

neither universally accessible nor equivalently used. For his purposes it is instead sufficient

36

91 Yochai Benkler, The Wealth of Networks: How Social Production Transforms Markets And Freedom (Yale University Press, 2006), http://www.benkler.org/Benkler_Wealth_Of_Networks.pdf/.92 Jurgen Habermas, The Structural Transformation of the Public Sphere: An Inquiry Into a Category of Bourgeois Society (MIT Press, 1991), 85. 93 Nancy Fraser, “Rethinking the Public Sphere: A Contribution to the Critique of Actually Existing Democracy,” in Habermas and the Public Sphere, Craig Calhoun ed (MIT Press, 1992), 109-142.

that the networked public sphere a) offers “universal intake” in principle and b) has a broader

“intake basin” in practice than that of the mass-media.94

Second, as Ben Roberts observes, it’s “very important for Habermas that [the public

sphere] for rational debate is kept separate from the political apparatus.”95 According to Mark

Warren, “Habermas emphasizes that public spheres cannot be organizers of collective action,

and must be protected from imperatives of collective action...[all democratic organizations]

must distinguish arenas of decisions and organizations of action.”96 For Benkler, however, the

promise of the networked public sphere is precisely the opportunities it offers for the two to

mix: most his examples, such as the 2004 boycott of Sinclair Broadcasting Group organized

through progressive blogs, are cases in which the processes of establishing and enacting a

consensus blend together into some more potent political accelerant than either on their own.

Third, as Stuart Geiger notes, the function of the public sphere most critical for

Habermas is “the admittedly arduous process of negotiating with those that one would not

ordinarily meet in private life.”97 For Habermas, democratic solidarity arises not from merely

mixing people in space, but from interlocuters conferring equal standing and respect upon each

other, the first step in a long process of exhaustive controversy through which different

lifeworlds are made commensurate. Geiger contends that the very painlessness of distributed

filtration mechanisms precludes them from functioning as Habermasian public spheres. But

37

94 Benkler, Wealth of Networks, 182, 236, 260. Howevr, Eszter Hargittai has suggested that Benkler may not have met these conditions; Hargittai, “Whose Networks? Whose Wealth?” Crooked Timber, May 30, 2006, http://crookedtimber.org/2006/05/30/whose-networks-whose-wealth/. 95 Ben Roberts, “Beyond the ‘networked public sphere’: Politics, participation and technics in web 2.0.” The Fibreculture Journal 14 (2009), http://fourteen.fibreculturejournal.org/fcj-093-beyond-the-networked-public-sphere-politics-participation-and-technics-in-web-2-0/.96 Mark Warren, “The self in discursive democracy,” in The Cambridge Companion to Habermas (Cambridge University Press, 1995), ed. Stephen White, 171-172; cited in Roberts, “Beyond the ‘networked public sphere.’” 97 R. Stuart Geiger, “Does Habermas understand the Internet? The algorithmic construction of the blogo/public sphere.” Gnovis: A Journal of Communication, Culture, and Technology 1 (2009): 1-29, http://gnovisjournal.org/2009/12/22/does-habermas-understand-internet-algorithmic-construction-blogopublic-sphere/.

Benkler defines the public sphere as “the set of practices that members of a society use to

communicate about matters they understand to be of public concern and that potentially

require collective action or recognition.” In other words, for Benkler the public sphere doesn’t

develop between people: it emerges across the network.98 Indeed, he implies that he values the

networked public sphere precisely for its peaceable synthesis of public opinion through a

liberal pluralist process as against the critical deliberative tradition of Habermas which seeks to

negotiate common understanding.99

The cumulative effect of these various departures is that Benkler transforms public

spheres from a) universally accessible, politically neutral discursive spaces in which people

negotiate mutually intelligible lifeworlds through communicative action to b) comparatively

accessible, politically engaged speech systems configured by technology and policy through

which people are more or less free to engage in speech acts. Benkler’s use of the term ‘public

sphere’ has led most to construe his as a Habermasian project, but his contribution is actually

much more innovative and interesting than simply extending Habermas’ concepts around a

new domain, for instead he has developed his own ‘digitally native’ model of the public sphere

which lends itself to an entirely different mode of investigating the Internet.

Because Benkler’s model shifts the purpose of the public sphere from negotiating

mutually intelligible lifeworlds to providing freer platforms for speech, he values the

networked public sphere because he believes it more autonomous than the mass-mediated.100

He characterizes the mass-media as a narrow band of speakers or subjects filtered to high

visibility by state or market power and the Internet as a broad band of speakers or subjects

38

98 Benkler, Wealth of Networks, 177. 99 Benkler, Wealth of Networks, 184.100 Benkler, Wealth of Networks, 16.

filtered to high visibility by mechanisms broadly distributed to users. The economics of

networked information production shift the most relevant considerations for freedom from

publishing to filtering: when anyone can speak, “the central point of failure becomes the

capacity to be heard—who listens to whom, and how that question is decided.”101

It is critical to understand that the Internet does not remove this point of failure. Indeed,

it cannot be removed, for, as in the Babel objection, if anyone can speak, no one can be heard.

Instead, Benkler argues that the Internet diffuses and distributes it. Unlike the mass-media, he

argues, the networked public sphere possesses “no single point of failure for discourse: no

single point where observations can be squelched or attention commanded—by fiat or with the

application of money.”102 According to Benkler, the well-documented unequal distributions of

attention to websites are not the product of power, but rather “emerge from many small-scale,

independent choices where free choice exists...power law distributions of attention to Web sites

result from random distributions of interests, not from formal or practical bottlenecks that

cannot be worked around...”103 These random distributions are collected, computed, and

reflected, he argues, not by emperors or editors but by distributed filtering mechanisms. Thus,

“we see the Babel objection solved on a distributed model, without anyone exerting formal

legal control or practical economic power [through the] coordinate effects [of] uncoordinated

actions.”104

The crux of Benkler’s argument is thus that the networked public sphere has “no

obvious points of control or exertion of influence—either by fiat or by purchase.”105 From this

39

101 Benkler, Wealth of Networks, 238.102 Benkler, Wealth of Networks, 271.103 Benkler, Wealth of Networks, 173.104 Benkler, Wealth of Networks, 173; 5.105 Benkler, Wealth of Networks, 177.

perspective the Digg Patriots can be seen to disrupt the networked public sphere by

identifying and seizing upon a third point of control, exerting their power not through the state,

or through the market, but through the algorithm. Under their sway Digg reflected not solely the

coordinate effects of uncoordinated actions, but also the coordinate effects of coordinated actions. As

such, the Patriots, aligned with their mutuals and empowered by algorithms, threaten to

reintroduce what Benkler calls “heteronomy, of dependence on the judgment of others that

subjects individuals to their control...”106 Suddenly it is no longer clear how or by whom the

power law distributions of attention are produced! This is a problem, because for Benkler

there is a sharp normative distinction between unequal distributions of attention produced

from uncoordinated action (good) and those produced by coordinated action (bad).

Like collective intelligence, the networked public sphere can be thus disrupted by

coordinated activity, but unlike Surowiecki, Benkler embraces situated human action, because

“liberal theories that ignore culture are rendered incapable of answering some questions that

arise in the real world and have real implications for individuals and polities.”107 Instead,

Benkler advocates developing an approach “able to diagnose different conditions in the

practical cultural life of a society as more or less attractive from the perspective of liberal

political theory.”108 From this perspective the Patriots can be seen as problematic culture

warriors who produce normatively bad cultural conditions by sidestepping the synthesis of the

networked public sphere, refusing to either peaceably clear disputes or negotiate mutually

intelligible lifeworlds. For Benkler, there needs to be some form of discourse, even if it is

emergent across a network rather than between individuals. The Patriots, however, sought to

40

106 Benkler, Wealth of Networks, 173.107 Benkler, Wealth of Networks, 280.108 Benkler, Wealth of Networks, 280.

simply suppress the ‘other side’ rather than engage it in deliberation, even going so far as to

treat comments instrumentally rather than discursively. Instead of clearing their disputes or

negotiating lifeworlds, they worked to escalate disputes and totalize lifeworlds.

Compared to collective intelligence, the networked public sphere offers a much more

theoretically sophisticated model by which to understand and assess user-generated

censorship. It provides a clear vision for what the Internet ought to be (a freer platform for

speech) and a simple rubric by which to evaluate the effects of the Patriots against its

realization (by reintroducing heteronomy and refusing to engage in the synthesis of opinion).

However, the networked public sphere model also doesn’t completely cover some of the

core characteristics of user-generated censorship. While Benkler is very concerned with power

wielded by traditional authorities, he doesn’t provide a conceptual apparatus to think about the

problems posed by algorithmically-empowered individuals. There’s also a huge liberal theory

foundation beneath Benkler’s work - the focus on speech acts, on freedom, on abstract

individuals, and so forth - that may not be helpful to import to discussions of user-generated

censorship. And the peaceable vision of the networked public sphere doesn’t quite seem to

capture the fighting spirit of the Patriots, who did not think of themselves as engaged in the

emergence of a consensus, but sought instead to win a contest and simply change everyone else.

As with collective intelligence, this doesn’t mean that the networked public sphere is a

bad model, just that it might not be the useful model by which to understand user-generated

censorship specifically. An even more effective model might account for factors that both of

these miss: the core coordination, the argumentative style, the desire to effect a sort of change

through the assembled strength of many actors. One such model might be drawn from the field

of science-technology studies, particularly the work of Bruno Latour.

41

3. Or: As Actor-Networks Assembling Artifacts

Latour is a French sociologist perhaps best known for studying the means by which

scientific arguments are made stronger, transforming mere propositions into hard facts. Yet

Latour doesn’t fit comfortably in either camp of the so-called ‘science wars’ of which he is a

indisputably a veteran: as Graham Harman observes, Latour has been “attacked

simultaneously for opposite reasons,” by scientists as “another soft French relativist” and by

social theorists as “a sellout to fossilized classical realism.”109

One of Latour’s central theses may be oversimplified as follows: all facts (arguments,

ideas, institutions, and so on) are socially constructed, but they are not all constructed equally

strongly, nor are they constructed solely by and of humans. In order to study or explain a fact

(of what and by whom it is built; why and to where it spreads), the social scientist must not

appeal to generalized ‘social forces’ which ‘lie behind,’ ‘contextualize,’ or ‘give power to’ the

argument. Instead, they should trace the specific human and nonhuman actors who are

recruited to form the fact. “I can now state the aim of this sociology of associations more

precisely,” writes Latour: “there is no society, no social realm, and no social ties, but there exist

translations between mediators that may generate traceable associations.”110

For example, consider a researcher who claims to have discovered a new bacterium.

When challenged to defend her claim, she may cite the microscope through which she saw it,

the method through which she defined it, and another journal article which guided her work.

The microscope, the method, and the article are all actors which she enrolls into an alliance

42

109 Graham Harman, Prince of Networks: Bruno Latour and Metaphysics (re.press, 2009), http://www.re-press.org/book-files/OA_Version_780980544060_Prince_of_Networks.pdf, 5. 110 Bruno Latour, Reassembling the Social: An Introduction to Actor-Network-Theory (Oxford University Press, 2005), http://dss-edit.com/plu/Latour_Reassembling.pdf, 108.

supporting her claim, and her claim depends on their continued cooperation. If her microscope

is well-calibrated, her method sound, and the article credible, then her allies held fast, and her

claim stays strong. But if the lens is defective, her method obsolete, or the article flawed, they

betrayed her, and her claim may fail. As Harman writes, from a Latourian perspective all facts

are fundamentally fights, and each brawler only as strong or as weak as its assembled allies:

It is never the actant in naked purity that possesses force but only the actant involved in its ramshackle associations with others, which collapse if these associations are not lovingly or brutally maintained...Any actant has a chance to win or lose, though some have more weaponry at their disposal. The loser is the one who failed to assemble enough human, natural, artificial, logical, and inanimate allies to stake a claim to victory.111

But while Latour can be reasonably said to believe in the ‘social construction of

scientific facts,’ for him the ‘social’ refers not to some exclusive domain or character of human

sociality, but the process of associating together many human and nonhuman actors. Facts,

propositions, ideas, concepts are not more or less ‘objectively true,’ but more or less well-

constructed. The strength or weakness of a fact arises, not from the fact itself, but from the

assembled strength or weakness of its allies, many of which are nonhuman. From this

perspective, atoms are more real than angels not because the former exist in reality and the

latter in prayer, but because atoms are more strongly linked with many other actors, like

mathematics and bombs.112

Latour thus lends social constructions the stubborn strength of nonhuman actors.

Anyone can dispute Newtonian physics, but they must battle both the Royal Academy of

Sciences and also apples insistently falling to earth. Against such tightly-knit facts the gods

43

111 Harman, Prince of Networks, 20. 112 Cf Harman, Prince of Networks, 110.

themselves contend in vain unless, like a Swiss patent clerk, they can convince enough

assembled actors to turn traitor such that it unravels itself like an old sweater.113 The social

scientist thus investigates a fact which she wishes to understand as an engineer investigates a

mountain through which she wishes to tunnel: by “[negotiating] with the mountain at each

stage of the project, testing to see where the rock resists and where it yields, and [being] quite

often surprised by the behavior of the rock.”114

Actor-networks become stronger by incorporating more and stronger actors into

themselves. The more associations the more durable the actor-network becomes because each

member has a stake in the whole thing holding together. The actor-network acts by translating

and transforming along the chain: “every time you want to know what a nonhuman does,” writes

one of Latour’s alter egos, “simply imagine what other humans or other nonhumans would

have to do were this character not present.”115 These actors became delegates of the actor-

network which issue forth like Nazgûl from Mordor. But no delegate is neutral, and no

delegation is costless. A hotel wishing to greet its guests with open doors may hire a porter, or

it may install a machine; each may solve the problem, but only after having been disciplined in

particular ways with different effects.

The most reliable delegates may congeal into black boxes. Like utterly trusted

lieutenants, black boxes seem to carry out orders without imposing their own biases, opinions,

motivations, or other transformative, translative effects. But this is only an illusion, for any

black box may be opened to reveal its politics, either through the careful prying of a

determined investigator or after being exploded by the box’s own unexpected behavior.

44

113 Cf Harman, Prince of Networks, 22. 114 Harman, Prince of Networks, 18. 115 Jim Johnson, “Mixing humans and nonhumans together: The sociology of a door-closer.” Social Problems (1988): 298-310; 299.

Harman has compared Latour’s black boxes to Heidegger’s tools: taken-for-granted and ready-

at-hand until they break or malfunction, at which point their contingencies, dependencies, and

configurations suddenly leap screaming to the foreground.116

From this perspective we might see the Patriots as assembling an actor-network

through Digg. Like the Pasteurization of France, the Patriotization of Digg proceeded by and

through the enrollment of mutuals as delegates, evaluated on their technical characteristics and

disciplined like door-closers. The most reliable mutuals came to be seen as black boxes, but as

soon as they failed the alliance by behaving unexpectedly or unreliably they were inspected

and discarded like broken hammers. Algorithms could be allies too: like a megaphone which

amplifies human voice, their strategic leveraging allowed the Patriots to make their points more

loudly. Yet, like unreliable mutuals, those same algorithms could also seem to work against the

Patriots, weaken their arguments, and leave them cursing the human and nonhuman traitors

which had betrayed their alliance.

In this chapter I have not conducted a proper actor-network analysis of the Digg

Patriots. However, I find the conceptual toolkit of actor-network theory, and more generally

science-technology studies, to offer several advantages as an approach for thinking about user-

generated censorship. First, this approach enables the descriptive study of user-generated

censorship without being weighed down by the normative baggage of other models. Second, it

begins from a presumption of coordinated activity, and provides a versatile set of tools by

45

116 Harman, Prince of Networks, 34. Indeed, it transpires that there are sufficient connections to merit a Actor-Network Theory/Heidegger mailing list called ANTHEM based out of LSE: https://groups.google.com/forum/?fromgroups#!forum/anthem-lse. ANTHEM hosted a conference and discussion between Harman and Latour after the publication of Prince of Networks, the transcript of which was itself turned into a book. See Erdelyi et al, The Prince and the Wolf: Latour and Harman at the LSE (John Hunt Publishing, 2011).

which to track and understand the alliances of humans and nonhumans which emerge through

sociotechnical systems.

But if we are to see the Patriots as making arguments through Digg, we need not only

ask how these arguments are made (that is, by the association of human and nonhuman

actors), but also the form which these arguments eventually take. As Harman writes, for Latour

“truth and reality are assembled through chains of actors in the same way that bills go through

Congress: slightly transformed and translated at each step, and failing as often as they succeed.

All reality is political, but not all politics is human.”117 Truth and reality are made manifest in

things. A hypothesis becomes a finding, becomes a result, becomes a fact. A discussion becomes

a bill, becomes a law, becomes a right. But into what does user-generated censorship harden?

What are the artifacts of its politics?

IV: THE SOCIAL PRODUCTION OF WHAT?Take the library of the British museum, for instance, valuable and useful

and accessible as it is: what chance has a work of being known to be there, merely because it is there? If it be wanted, it can be asked for; but to be

wanted it must be known. No one can rummage the library.- Augustus De Morgan, to Charles Babbage118

1. The Ark in the Archives: On the Centrality of Sociotechnical Systems

In asking after the artifacts of the politics of user-generated censorship, I do not mean,

as Langdon Winner has famously argued, artifacts which embody a particular politics, but

46

117 Harman, Prince of Networks, 89. 118 Quoted in Gleick, The Information, 410.

merely artifacts which are the result of a politics.119 If, as I have argued, alliances of actors

assemble together the front page of Digg, the Facebook News Feed, Google Search results,

and so on through political processes, then these assemblies, once assembled, are themselves

the artifacts of politics. But what are these artifacts, and what are their metaphysics? When the

Patriots coordinate to change Digg, what are they changing, and how does it change?

It’s trivial to observe that social media are socially produced; the far more interesting

question is what is produced through the process of social production.120 One possible answer is

a kind of representation, like a map, which guides users through an otherwise unnavigable

flood of information. Such signposts are so central as to be invisible, yet profoundly influential

despite - indeed, because of - their invisibility. “Information infrastructure is a tricky thing to

analyze,” observe Bowker and Star, because “Good, usable systems disappear almost by

definition. The easier they are to use, the harder they are to see.”121

Perhaps an example, set in a familiar scene, can make them more visible. In the closing

moments of Raiders of the Lost Ark, shortly after federal agents refuse to reveal its location to

Indiana Jones, the camera cuts to anonymous hands nailing the ark inside a wooden crate.122

An old man wheels it slowly down the aisle of an enormous warehouse packed full of

47

119 Contra Winner, “Do Artifacts Have Politics?”, compare Bernward Joerges, “Do Politics Have Artefacts?”, Social Studies of Science 29, no. 3 (1999): 411-431. Whereas Winner believes that artifacts embody a politics, Joerges believes that they usually cannot, and argues compellingly that Winner’s most famous example - Robert Moses’ bridges - is simply incorrect. Instead, Joerges advocates studying not the artifacts, but the processes which produce them. By asking after the artifacts of a politics I am somewhere between the two: interested in the things produced by the processes Joerges cares about without requiring a specific politics as Winner would. I elaborate on my interest in the third section of this chapter and in the conclusion. My quip about the result, not the reason, is a riff on a legal realist catchphrase: see generally Karl Llewellyn, The Bramble Bush: On Our Law and Its Study (Oceana Publications, 1930); see also Elizabeth Mensch, “The History of Mainstream Legal Thought,” in The Politics of Law: A Progressive Critique, ed. David Kairys (Basic Books, 1998), 23-53. 120 This observation, along with the title of this chapter, is riff on the work of (and with apologies to) Ian Hacking. Hacking, The Social Construction of What? (Harvard University Press, 1999). 121 Bowker & Star, Sorting Things Out: Classification And Its Consequences (MIT Press, 2000), 13. 122 Raiders of the Lost Ark. Directed by Steven Spielberg (1981).

innumerable crates, each indistinguishable from that which contains the ark. The worker turns

left. As ominous music builds, the ark vanishes from view, effectively lost in plain sight among

a sea of infinite, identical crates.

Suppose Indiana Jones decided to search for this ancient and dangerous artifact. It is

not enough for him to merely infiltrate the warehouse. He needs to know where to begin

looking, yet despite the considerable determination and unorthodox methods of Professor

Jones, his search would almost certainly be stymied without the aid of some Ariadne’s thread

to follow through the warehouse. Blair and Stallybrass have traced the origins of the verb “to

file” to mean the literal stringing of documents along a wire - that is, a filum - and so too must

Jones move only ever down the organizational filaments leading to the ark.123 Thumbing

through the folder of “O” for “Occult” he may find a slip of paper indexing its location; by the

door, a map depicts the layout of the warehouse, which he may navigate by following signs

amidst the stacks. Eventually Jones may indeed find the crate, but only if every step along the

way holds true. If the index is off, the map incorrect, or any other link in the chain snipped (by

earnest error or strategic mislabeling), Jones is lost, and so too the ark. No one can rummage

the warehouse.

This story illustrates how objects may be formally accessible yet contingently findable. It

underscores the central role of sociotechnical systems in making available information; that is,

assisting its location or discovery. Indeed, to archive the ark in the warehouse means not only to

put the box on a shelf but also to enroll it in a system through which it may later be located.

Stamping with an identifying number, indexing into an inventory, mapping against the stacks:

each nonhuman actor in this chain refers to the next, paving the pathways along which Jones

48

123 Blair & Stallybrass, “Mediating Information 1450-1800,” This Is Enlightenment, ed. Siskin & Warner (Chicago: University of Chicago Press, 2010), 152.

must travel. The spacious warehouse in Raiders collapses in practice to these tiny, two-

dimensional conduits through which knowledge actually circulates. The ark is ‘in’ the archives,

but effectively can’t be found ‘there’; instead, it’s found through sprawling sociotechnical

systems. As Latour once said at a lecture on globalization, “…we always tend to exaggerate the

extent to which we access this global sphere…There is no access to the global for the simple

reason that you always move from one place to the next through narrow corridors without ever

being outside.”124

Discussions of censorship usually focus on an imagined formal accessibility of objects: is

the book/painting/record ‘in’ the library/gallery/store? But once we see objects as existing not

‘in’ space but through passageways, our focus reorients around these newly-realized relations.125

From this perspective the questions of how, by whom, and for what reasons these corridors are

tunneled and reinforced elevates from ancillary interest to central significance. User-generated

censorship operates by reconfiguring these corridors with the assistance of social media

machines, a metaphor I’ve adapted from the media historian Markus Krajewski.

2. Social Media as Machines which Manage the Flood

With the invention and spread of printing with movable type, a complaint arises in the learned reading world. It is the book flood, always a nautical or irrigation metaphor, that has a disturbing effect on readers in the newly established privacy of their studies...Only

49

124 Bruno Latour, “Spheres and networks: two ways to reinterpret globalization.” Harvard Design Magazine 30 (2009): 138-144. http://www.bruno-latour.fr/sites/default/files/115-SPACE-HARVARD-GB.pdf/.125 Indeed, as Stuart Geiger has reminded me, a strict ANT approach would suggest that there can in fact be no distinction since an object is always and only defined in association with its relations. See also Julie Cohen’s model of ‘embodied spatiality’: Cohen, “Cyberspace as/and Space.” Columbia Law Review (2007): 210-256; Chris Peterson, “The Ark in the Archives: Toward a Theory of Link-Oriented Antiepistemology,” MIT Center for Civic Media, November 6, 2012, http://civic.mit.edu/blog/petey/the-ark-in-the-archives-toward-a-theory-of-link-oriented-antiepistemology/.

when the library is inundated is the need to deal with all of this material recognized.126

So begins the second chapter of Krajewski’s improbably enthralling Paper Machines:

About Cards & Catalogs, 1548-1929, which traces the development of card catalogs as a response

to newly abundant information after the introduction of the printing press. Krajewski’s book is

a refreshing reminder that neither our contemporary concerns nor favored metaphors for

information overload are new. “...Paper also became so cheap, and printers so numerous, that

a deluge of Authors covered the land,” complained Alexander Pope in The Dunciad. As James

Gleick notes, “Deluge became a common metaphor for people describing information surfeit.

There is a sensation of drowning: information as a rising, churning flood.”127 The deadly flood

had to be tamed by tools: “...the shortness of our life and the multitude of things that one must

know today to count among the learned do not allow us to do everything ourselves,” wrote

Gabriel Naude in 1627 of the need to rely on instruments to manage the book flood.128

Card catalogs were among the instruments developed to help readers navigate this

intensely-felt flood of information. While the earliest bibliographic practices indeed consisted

of little more than lists of books in a library’s collection, it was Leibniz, as director of the

library at Wolfenbüttel, whom Krajewski credits with first arriving at the idea of cataloging

books on paper slips which could be easily rearranged to account for new additions.129 The

first actual catalog wasn’t created for nearly a century, when one was developed for the

Viennese royal library, coincidentally during the same decade that the Empress Maria Theresa

ordered conscription numbers be publicly posted as addresses to help the monarchy identify

50

126 Markus Krajewski, Paper Machines: About Cards & Catalogs, 1548-1929 (MIT Press, 2011), 9. 127 Gleick, The Information, 384. 128 Ann Blair, Too Much To Know: Managing Scholarly Information Before The Modern Age (Yale University Press, 2010), 57. 129 Krajewski, Paper Machines, 22.

soldiers among “an indistinct sea of houses.”130 As their respective empires expanded, both

royals and readers developed methods to help them find objects amidst their newfound

abundance.

Krajewski calls card catalogs paper machines because they “can be described as a chained

mechanism” which does the work of organizing information when powered by the user’s

hand.131 Following the dominant metaphors of the time, he argues that the organizational work

which card catalogs do can be understood as channeling the book flood by digging canals

which tame the torrent into more manageable streams. But there is no translation without

transformation: card catalogs, like all mediators, do not simply transmit information

unperturbed, but impose their effects as they go. Krajewski notes that “it would be insufficient

to mark only beginning and end, sender and receiver. Rather, what is crucial is the way this

transfer occurs, including any disturbances, changes, stoppages, irritations and detours,” and

so he carefully tracks the manner by which card catalogs transform the books to which they

also refer.132 Cards in catalogs, as mobile, rearrangeable pointers to books located in the the

stacks, also “[embody] a highly compressed data set that characterizes the book to be found...

[becoming] a representation of the text - which now need no longer be read every time.”133 A

card must compress, distort, transform the book to which it points; else it is not a card, but the

book itself, and the flood continues to rise.

The key insight here is that card catalogs play a central, rather than incidental or

ancillary, role in connecting readers to books. In any collection of meaningful size, card

catalogs mediate the two; it is only by traveling through cards that readers find books. Yet it is

51

130 Krajewski, Paper Machines, 21. 131 Krajewski, Paper Machines, 7.132 Krajewski, Paper Machines, 3.133 Krajewski, Paper Machines, 23.

precisely their role in making books more findable which makes them an attractive avenue of

intervention for those who wish to make books less findable. Krajewski calls the specter of lost

cards “the sum of all librarian’s fears,” and this specter came to haunt the University of Illinois

in the 1960s when, according to Nicholson Baker, student radicals targeted card catalogs for

disruptive action.134 Over 50,000 cards were burned and scattered in February 1969: “The

tragedy,” library Dean Robert Downs told the Daily Illini, was that now “there won’t be a

complete record of all the books in the stacks and in the various departmental libraries.”135 In

response, the University of Chicago placed its catalog under armed guard for months while it

carried out a secret project to microfilm a backup copy of its cards.136 This is the strange,

dialectical quality of finding devices: that, because they help find, they may also be repurposed

to inhibit finding.

Krajewski models card catalogs as machines which, through their mobility and

reconfigurability, assist readers in managing the book flood. The raging river of print never

receded, but was eventually tamed by a system of locks and canals, its well-channeled waters

eventually powering the turbines of knowledge production.137 We might extend his model to

understand social media as machines which help manage the digital flood. Nautical metaphors

for information abundance are as common today as they were in the time of Leibniz and Pope;

like our analog ancestors, we constantly feel at risk of drowning beneath its rising tides, and

need new mechanisms by which to channel the waters.

52

134 Krajewski, Paper Machines, 54; Nicholson Baker, The Size of Thoughts: Essays and Other Lumber (Vintage, 1997). 135 “Vandals Destroy About $15,000 In University Library Catalog Cards”, Daily Illini, May 24, 1969, page 8.136 Neil Radford, “Books On Film,” The Core: College Magazine of the University of Chicago, summer 2010, http://thecore.uchicago.edu/Summer2010/eye-books-on-film.shtml/.137 See also Krajewski’s discussion of the economic relations of card catalogs, e.g. quoting Goethe on libraries as “One feels as if in the presence of capital that noiselessly yields unpredictable interest” in Paper Machines at 24.

Like cards constituting compressed versions of books, the links exchanged through

social media pack whole websites, pictures, and films into tiny, mobile representations. A link

on Digg is not merely a pointer, but a “highly compressed data set that characterizes” the site

to be found at the other end, annotated by net upvotes and user comments. But these links are

mobile vertically in a way that cards are not. Social media streams are usually organized around

metaphors of depth: sites like Digg and reddit enable users to vote good content “up” and bad

content “down”, while services like Twitter and Facebook algorithmically curate “top” posts.

It’s as if the information flood acquired a new digital dimension; as if, when information

overflowed its banks, social media responded by digging the riverbed deeper. Dave

Weinberger has argued that, on social media, “[filters] no longer filter out. They filter forward,

bringing their results to the front.”138 He might instead have said, with a bit more accuracy,

that social media don’t filter out: they filter down.

This is a fundamentally different model of knowledge management than paper

machines. Card catalogs often have outdated ontologies but within their schema all books in a

given collection are equivalently available; nothing is ever at the “bottom” of a card catalog.

But the users of social media, by operating the chained mechanisms of algorithms, surface some

stories, submerge others, and stop still more at their source. Reading the front page of Digg, or

the Facebook News Feed, is like standing atop a rock in a dark, swift river: while some objects

are clearly visible, having been buoyed to the surface, countless others rush by invisibly in the

depths. The latter can be found by a determined reader - but only if she dons goggles, hold her

breath, and dives deeply in search of them, and even then only if she knows to look.

53

138 David Weinberger, Too Big to Know: Rethinking Knowledge Now That the Facts Aren't the Facts, Experts Are Everywhere, and the Smartest Person in the Room Is the Room (Basic Books, 2012), 11.

Like card catalogs, social media sit centrally between readers and information; their

mechanisms make their compressed referents to content more or less visible. John Rajchman,

building on Foucault, argues that “spaces are designed to make things seeable, and seeable in a

specific way.”139 So too with social media which, with their surfaces and submergences, are

also designed to make certain things unseeable in a specific way. Information is always

immediately at risk of being drowned beneath the waterline in social media streams,

submerged sufficiently to suffocate for lack of air and attention. Taina Bucher has called this

the “threat of invisibility:” unlike the all-seeing eye of the panopticon which regulates by

imposing visibility before its inescapable gaze, the mechanisms of social media regulate by

imposing “the constant possibility of disappearing” beneath the surface.140

User-generated censorship enforces this threat, leveraging social media machines to

make certain information effectively unfindable. Like student radicals sinking whole sections of

the stacks back into the book flood, the Digg Patriots submerged liberal links deep in social

media streams, rendering them less visible than they would have been without their

intervention, sucked under by an algorithmic riptide. These objects remain accessible but

suddenly less available, the long chain of references linking them buried too deeply to be easily

followed by casual travelers. Whole pathways drown beneath the waves like lost continents,

while others pop to the surface in their stead, distracting from what has gone missing.

The subversive genius of user-generated censorship is that it intervenes not on objects

but on the routes through which they can be found, which are erased or made to appear less

interesting to travel down. It’s a quiet, subtle mode of strategic action, which, through the

54

139 John Rajchman, “Foucault’s art of seeing,” October 44 (1988): 88-117, http://www.jstor.org/stable/778976. Cited in Taina Bucher, “Want to be on the top? Algorithmic power and the threat of invisibility on Facebook." New Media & Society 14, no. 7 (2012): 1164-1180.140 Bucher, “Want to be on the top?”

laundering of algorithms, is often not even distinguishable as such. Censoring objects is usually

loud and noisy: a banned book is obviously gone, and a blocked website is self-evidently

inaccessible. User-generated censorship eschews such crude methods, instead hiding its targets

in plain, unseeable sight, sunk in the muck at the bottom of the social river.

3. The Social Production of Persuasive Panoramas

User-generated censorship is a strategic repurposing of the general filtering function

which tames the flood. By making information more or less available, social media produce

influential representations of the things which exist to be viewed, experienced, accessed, as

well as links leading directly to them. In other words, they make maps, providing both a list of

destinations and well-paved pathways through which they can be found. But, as Alfred

Korzybski observed, the map is not the territory.141 Indeed, the map cannot be the territory, for

a map of the earth the size of the earth would be just as unnavigable.142 The interesting

question, then, is not whether a given representation of the world is incomplete, but in what ways

it is incomplete; that is, which aspects were omitted in compression, with what effects?

In Reassembling the Social Latour discusses two different types of representations:

“clamps,” as he calls them, to help compress the world so it can be mapped. The first kind of

map Latour calls oligoptica.143 Unlike their implicit opposites, all-seeing panoptica, oligoptica

55

141 Alfred Korzybski, Science And Sanity: An Introduction to Non-Aristotelian Systems and General Semantics (Institute of General Semantics, 1958), 750. 142 Borges: “In that Empire, the Art of Cartography attained such Perfection that the map of a single Province occupied the entirety of a City, and the map of the Empire, the entirety of a Province. In time, those Unconscionable Maps no longer satisfied, and the Cartographers Guilds struck a Map of the Empire whose size was that of the Empire, and which coincided point for point with it. The following Generations, who were not so fond of the Study of Cartography as their Forebears had been, saw that that vast Map was Useless, and not without some Pitilessness was it, that they delivered it up to the Inclemencies of Sun and Winters.” Jorge Borges, “On Exactitude in Science,” Collected Fictions, translated by Andrew Hurley (Penguin Books, 1999), 325.143 Latour, Reassembling the Social, 181.

provide self-evidently narrow, limited lines of sight to the things surveyed. Deploying his

favored military metaphors, Latour characterizes a war room as an oligopticon, from which

generals may view scenes from a distant front only long as the attenuated connections hold; as

such, they are hyperaware of their dependence on satellites and radios, and that they are not

able to see what’s happening where the camera isn’t. The second kind of map Latour calls

panoramas. “Contrary to oligoptica,” he writes, “panoramas...see everything. But they also see

nothing since they simply show an image painted (or projected) on the tiny wall of a room fully

closed to the outside.”144 Panoramas, in other words, are representations which do not reveal

their own incompleteness. As Latour writes:

[Panoramas] are being painted every time a newspaper editorialist reviews with authority the ‘whole situation’; when a book retells the origins of the world from the Big Bang to President Bush; when a social theory textbook provides a bird’s eye view of modernity; when the CEO of some big company gathers his shareholders; when some famous scientist summarizes for the benefit of the public ‘the present state of science’; when a militant explains to her cellmates the ‘long history of exploitation’...They design a picture which has no gap in it, giving the spectator the powerful impression of being fully immersed in the real world without any artificial mediations or costly flows of information leading from or to the outside. Whereas oligoptica are constantly revealing the fragility of their connections and their lack of control on what is left in between their networks, panoramas give the impression of complete control over what is being surveyed, even though they are partially blind and that nothing enters or leaves their walls except interested or baffled spectators. To confuse them with oligoptica would be like confusing a war episode monitored from the U.S. Army war room in Tampa, Florida, with the same one related on Fox News when a retired general is commenting on the ‘day at the front’. The first account, which is a realist one, knows painfully well that it can become unreal as soon as communications are cut off; the second one might be just as real but it has a smaller chance of telling us whether or not it’s fiction.145

56

144 Latour, Reassembling the Social, 187.145 Latour, Reassembling the Social, 188.

From this perspective we might see the maps of social media as constituting persuasive

panoramas. Panoramas, because the pictures they provide do not volunteer their own

incompleteness; persuasive, because their reflexive silence fails to betray the many ways in

which their representations are incomplete, luring the viewer into believing they are looking at

a comprehensive and natural image of the world, as if through a clear lens, when in fact they

are watching a wholly political performance. Social media, as sociotechnical systems, can be

understood to assemble users and algorithms together to make arguments about what is

important, what is relevant, what is “trending,” what is “top,” and so on; these arguments are

manifest in the panoramic artifacts projected onto their front page, their News Feed, their

search results, and so on.

All representations of the world are compressed, but they are not all compressed the

same way, and the composition of representations is a primary concern in media studies. Critical

theorists across fields and activists of all stripes routinely critique different pictures of the

world as dominant and hegemonic while working to construct alternate and resistive narratives

from subaltern perspectives, and so on. Indeed, a core insight of media studies is that the

narratives, perspectives, and representations made through media help constitute our

ontologies, our ontologies our lifeworlds, and our lifeworlds our politics. What these Latourian

concepts provide is a way to distinguish between representations based on their epistemological

earnestness. Oligoptica see the world through tunnel vision, and continuously destabilize

themselves by foregrounding their own dependencies and contingencies. Panoramas, on the

other hand, are born stable, whole, and totalizing; they must be forced to admit their

incompleteness.

57

The power of user-generated censorship derives from these epistemological tensions.146

Panoramas are persuasive not because there is no gap in them but that there seems to be no gap.

Any absence is utterly inconspicuous: the eye seamlessly elides what it does not expect to not

see, like Muggles simply failing to notice the Knight Bus.147 Fundamentally, it’s a problem of

unknown-unknowns: there’s no real reason to look for what isn’t included in Google’s search

results, what isn’t highlighted in the Facebook News Feed, what doesn’t rise to the top of reddit

or Digg. The panorama provides users a convincing map of what exists and what is interesting

such that they do not notice what they do not notice.

Latour has recently shifted his focus from critique to composition, taking a surprising

Habermasian turn towards negotiating a common world.148 “A world that you have to compose

is not the same thing as a world you have to discover, and not the same thing as a world you

have to uncover,” says Latour; instead, it is something you make.149 One way to understand

user-generated censorship is as recomposing the common world produced by social media,

shifting objects between the foreground and background of the panorama such that the artifact

remains apparently natural and undisturbed. It’s as if a clever Church cleric, ordered to paint

fig leaves on nudes but aware of the attention his painfully obvious censorship might bring,

58

146 Or anti-epistemological: cf. Peter Galison, “Removing Knowledge”, Critical Inquiry 31, no. 1 (2004): 229-243.147 “‘How come the Muggles don’t hear the bus?’ said Harry. ‘Them!’ said Stan contemptuously. ‘Don’ listen properly, do they? Don’ look properly either. Never notice nuffink, they don’.’” - J.K Rowling, The Prisoner of Azkaban (Scholastic, 1999), 36. As you may have (dare I say it?) noticed, I have a bit more sympathy for how hard it can be to notice things than Mr. Shunpike. 148 Bruno Latour, “An Attempt At A ‘Compositionist Manifesto’”, New Literary History 41, no. 3 (2010): 471-490, available at http://www.bruno-latour.fr/sites/default/files/120-NLH-GB.pdf. See also “Prof. Bruno Latour 'From Critique to Composition'”: "And here compromise means etymologically promise together. Is there a way to promise together a common world which is composable? A world which does not exist but which can be made to be? Which is completely different from having the common world as a background known by matters of fact on which now you add the politics.” - 1:12:50, http://www.youtube.com/watch?v=-02aCvQ-HFs/.149 Latour, “From Critique to Composition,” at 30:00.

could instead, through some trick of the eye, subtly alter the painting such that viewers simply

elided the offending portions of their own unconscious accord.

At its most basic, the user-generated censorship theory of change is a classic technique

of agenda setting, of altering media representations, akin to convincing a newspaper editor to

not run a story or a textbook editor to change her history book. Yet it is not merely this, for

these are new levers, distributed to different actors, laundered by algorithms, emerging from

complex systems often inaccurately presumed to be free of such politics. Almost anyone, with

the timely aid of a few allies, can cut a backroom deal with an algorithm to help suppress a

story, change a narrative, shift a lifeworld, and these interventions leave far subtler traces than

their analog analogues. No books are burnt. No redactions highlighted in black marker. Just

some roads quietly dropped from the map, pathways effaced or erased, such that their

destinations become much harder to find. “Can’t get there from here,” my grandparents from

Maine sometimes simply say; so too it is with user-generated censorship.

V: CONCLUSION

...[Social media practices] reflect something genuinely new, and as yet not clearly theorized, distinct equally from Habermasian communal

conversation-as-deliberation as from the blandly managerial product, shaped by layers of human talent for the broadest possible distribution, of

Adorno and Horkheimer's Kulturindustrie. - Finn Brunton, “Spam: A Shadow History of the Internet”150

1. The Relevance of User-Generated Censorship

In this thesis I have studied user-generated censorship as an emergent mode of

intervention made possible by social media. I grounded it in the case of the Digg Patriots, who

assembled and maintained human and nonhuman allies in order to make Digg different than it

59

150 Brunton, Spam, 165-166.

would have been without their intervention. I compared social media machines to paper

machines as sociotechnical systems which make specific information more or less available

amidst general abundance. Finally, I analogized the artifacts of social media machines to maps

which provide more or less obviously incomplete representations of the world, as panoramic

pictures which can be recomposed to alter the pathways paved by and through them. But an

important question remains: so what? What is the significance of user-generated censorship?

Why does it matter? What makes it worth thinking about, worth writing about - and, if you’ve

made it this far, worth reading about?

The most straightforward answer is that social media are politically important. They are

among the most influential informational intermediaries we have. To the extent that social

media mediate people - to the extent they pave the routes of passage through which we must

travel to come to know things, as well as lists of destinations which may come to be known -

changing the composition of social media changes the composition of the (un)common world of

which they are themselves a constitutive element. As such, the representations made by and

through social media merit the same critical attention long paid to other representations of the

world made by newspapers, television news, history books, and so on.

A corollary observation is that the artifacts of social media are no more ‘natural’ or

‘neutral’ than the artifacts of other media - nor could they be. It is not only, as Daston and Galison

have shown, that the very meaning of ‘objectivity’ has changed over time.151 Nor is it only that

there is no way to transform the territory into the map without losing something in the process

with political effects. More fundamentally, it is that the territory itself effectively does not exist

to be mapped, but is in fact constituted by its mapping. When James Surowiecki argues that

60

151 Daston & Galison, Objectivity (Zone Books, 2010).

for “Google searches, there is a definitive answer, which at some point is settled once and for

all...whether it found the right page or not,” he mistakes, to paraphrase the legal realists, the

reason for the result.152 Search is as indeterminate as law: the ‘right page’ is not that which

preexists the search, to be uncovered by Google, but that which is produced by the search,

through sociotechnical alliances assembled precisely to make the proposition of ‘rightness’

seem more or less persuasive.153

Social media cannot be ‘natural’, yet we desperately want them to be, to the point of

creating truly bizarre ontological categories. Compare, for example, Google’s fascinating

prohibition of ‘unnatural links.’154 What makes authoring a hyperlink - which exists only ever

in an electronic world as produced by humans or their delegates - more or less ‘natural’? A

configuration of chromosomes? A state of mind? A mode of creation? Or consider The Atlantic’s

revelation that, in Facebook’s dispute resolution process, human moderators “optimize for half

a second” of time before deciding whether to retain or remove the flagged content based on a

complex set of fixed rules, an algorithm inscribed on paper rather than manifest in code.155

From the perspective of biology, Facebook’s moderators are unquestionably human, yet from

the perspective of the dispute resolution process, they are effectively a computer, one which

recovers the original definition as the job description of one who computes.156

61

152 Surowiecki, Wisdom of Crowds, 236. 153 On the indeterminacy of search, see James Grimmelmann, “The Google Dilemma.” New York Law School Law Review (2009): 939, http://works.bepress.com/james_grimmelmann/19/.154 “Link Schemes”, Webmaster Tools Help, accessed May 24, 2013, http://support.google.com/webmasters/bin/answer.py?hl=en&answer=66356/.155 Emily Bazelon, “How to Stop the Bullies,” The Atlanic, February 20, 2013, http://www.theatlantic.com/magazine/archive/2013/03/how-to-stop-bullies/309217/.156 See, e.g., “Human computer,” Wikipedia: The Free Encyclopedia, accessed May 24, 2013, http://en.wikipedia.org/wiki/Human_computer/.

posts of the Patriots, humans and algorithms are often ontologically indistinguishable,

oscillating across the nature/culture boundary depending on whether they are behaving as

expected at any given time. When mutuals or algorithms ‘worked,’ they seemed neutral,

natural tools; when they ‘resisted,’ they were instead opened to find the people and politics

packed inside. The same is true of their opponents, like Olson, who first began to suspect the

existence of an organization like the Patriots when Digg did not behave as he expected. Oddly,

actual people seem to be invisible in social media - until things go poorly. It is only when the

systems do not behave as expected that the presence of people becomes apparent as the source

of the biases, inadequacies, and errors. Thus many of these platforms forbid ‘manipulating,’

‘altering,’ or doing ‘unnatural’ things, which necessarily imagines some ‘unmanipulated,’

‘unalterated,’ ‘natural’ paradise lost. Perhaps it is my childhood talking, but there’s something

strangely Catholic about this: as if when human motivations, behaviors, or incentives are

recognized in social media, the first instinct is to read them as contaminant; as sin.

It is out of this gap between the designer and the user, between the author and the

authorized, that the very possibility of user-generated censorship emerges, like a plant growing

from cracks in concrete. It is only because Digg was imagined to produce some ‘unaltered’

representation of the world that the activities of the Patriots could cause such concern. It is

only because Google imagines a certain kind of links as ‘natural’ that other kinds of links can

become unnatural. When I began discussing J30Strike with a former Facebook engineer she

interrupted to exclaim “ah, you’re talking about people misusing the spam button” only

because she could envision a different use which is not misuse. When I asked a respected

Internet scholar about my case studies he offhandedly characterized them as “exploitations” of

social media because “everyone knows they aren’t supposed to behave like that.” The central

62

conflict of the Digg Patriots is suddenly shown to not be between liberals and conservatives at

all, but between the expectations of the designers of platforms and the acts of their resistive

users, revealing the actual hierarchies animating ‘democratized’ media.157

User-generated censorship is not, as it may initially appear, a simple story of bad

humans and algorithmic levers, but instead a complex narrative which interweaves the agency

and expectations of many ontologically uncertain actors. It is best understood not as a

standalone practice but as one which coevolved interdependently with a lush material-semiotic

ecology.158 User-generated censorship is not only relevant politically, but productive

intellectually, because the process of investigating its emergence leads one to the unexpected

conditions which made its emergence possible. It thus transforms from a subject of inquiry to a

vehicle of inquiry, facilitating access to certain questions which it makes more visible and

interesting to ask, questions which pertain not only to the continued study of user-generated

censorship, but more generally to the algorithmic turn.159

One set of questions are ethnographic. For user-generated censorship, as for other

algorithmic phenomena, it may be asked: who are the actors involved? What are their

categories and concepts, their epistemologies and ontologies? How may their associations be

traced? Such questions are best answered by appealing to the actors themselves. These

investigations may proceed through traditional ethnography, such as the interviews Malte

Ziewitz has conducted of search engine optimization engineers, or those Nick Seaver is

63

157 This, I think, is the core insights of Evgeny Morozov’s critique of the Internet as “The Internet.” To the extent that we see “The Internet” as an assembled object made of unspoken expectations what it “ought” to be, it inhibits us from asking certain questions from perspectives which do not share those assumptions. Morozov, To Save Everything Click Here.158 For an excellent history of a practice coevolving in a sociotechnical ecology, see generally Brunton, Spam; see also Stuart Geiger, “The Lives of Bots,” in Wikipedia: A Critical Point of View, ed. Lovink & Tkacz (Amsterdam: Institute of Network Cultures, 2011), http://www.stuartgeiger.com/lives-of-bots-wikipedia-cpov.pdf/.159 For one definition of the algorithmic turn, see William Uricchio, “The algorithmic turn: Photosynth, augmented reality and the changing implications of the image.” Visual Studies 26, no. 1 (2011): 25-35.

conducting of music recommendation algorithm developers.160 Or, they may employ more

radical methods, such as Stuart Geiger’s technique of “trace ethnography” which enables him

to decode large-scale quantitative documentary data into nuanced qualitative accounts of the

human and nonhumans who animate Wikipedia, or Taina Bucher’s careful excavation of the

concepts and categories embedded in the algorithms of programmed publics.161

Ethnography is a powerful method for eliciting and describing actors, their categories,

and their associations. Yet it can also be limited, for a faithful ethnography, informed by its

subjects, is often blind in the same ways they are. This is especially true for the accounts of

actor-network theory. Latour cheerfully characterizes ANT as “myopic” precisely because of

its shortsighted perspective, which then gives rise to the problem of “plasma”:

namely, that which is not yet formatted, not yet measured, not yet socialized, not yet engaged in metrological chains, and not yet covered, surveyed, mobilized, or subjectified. How big is it? Take a map of London and imagine that the social world [traced] so far occupies no more room than a subway. The plasma would be the rest of London, all its buildings, inhabitants, climates, plants, cats, palaces, horse guards...If knowledge of the social is limited to the termite galleries in which we have been traveling, what do we know about what is outside? Not much.162

The problem of plasma is fundamentally this: when you follow the actors, trace their

associations, and elicit their categories, you are left with a highly-descriptive but narrow

version of the subject of study. We know something about the Digg Patriots from their unusual

qualitative corpus. But what of the rest of Digg? What of the other users and their artifacts?

64

160 See Malte Ziewitz, “Evaluation as Governance:The Practical Politics of Reviewing, Rating and Ranking on the Web,” http://ziewitz.org/projects/evaluation-as-governance.html; Nick Seaver, “Knowing Algorithms”, http://web.mit.edu/comm-forum/mit8/subs/abstracts.html#seaver/.161 For an introduction to trace ethnography, see Geiger & Ribes, “Trace Ethnography: Following Coordination through Documentary Practices” in Proceedings of the 44th Annual Hawaii International Conference on System Sciences (HICSS). http://dx.doi.org/10.1109/HICSS.2011.455; Taina Bucher, “The friendship assemblage: Investigating programmed sociality on Facebook”, Television & New Media (2012), http://dx.doi.org/10.1177/1527476412452800/.162 Latour, Reassembling the Social, 244; 242.

What of the 70 million people who monthly visit reddit? Or the billions of human and

nonhumans actors who assemble together Google’s search results? We cannot trace all their

associations, and even if we could, we would just be tangling ourselves in an ever more

complex web of threads without any ability to characterize what is being woven.

Ethnography may thus be complemented by archaeology: the process of comparing the

artifacts of social media. Instead of tracing the associations of the actors who assemble an

artifact, this approach holds the artifact thus assembled against another, revealing the

differences between them. These comparisons do not themselves ‘say anything,’ but the process

of making visible these differences opens pathways for inquiry about why and how they are

different, for what reasons and with what results.

This approach is particularly useful for the study of social media. If, as I have argued,

the interlocking levers of social media machines often produce panoramic representations of

the world, these deceptive projections must be made to admit their incompleteness with the aid

of diagnostic instruments. For example, my colleague Nathan Matias at the MIT Center for

Civic Media has helped develop the OpenGenderTracking toolkit, which he has used to reveal

some startling differences between the gender representation of authors in U.K. newspapers.163

Nate is now building an application called FollowBias which visualizes the gender distribution

of one’s Twitter feed, thus foregrounding a factor of its composition which would otherwise

easily sink to the background.164

If the methods of ethnography provide tools which help investigators trace the assembly

65

163 Matias & Evans, “Women's representation in media: readers preferences for online news revealed”, The Guardian: Datablog, October 23, 2012, http://www.guardian.co.uk/news/datablog/2012/oct/23/women-media-representation-online-news. The OpenGenderTracking toolkit is available at http://opengendertracking.org/blog/2013/01/08/Tracking-Gender-In-Online-News/. 164 FollowBias is presently available for installation at http://followbias.com/.

of artifacts, these archaeological instruments, like magnifying glasses or slides, help them notice

differences between artifacts already assembled, which provides a path forward not only for

scholarly inquiry but also for pragmatic intervention. It is only once someone notices that their

Twitter feed is comparatively male, or their subreddit comparatively liberal, that they can

decide what they want to do about it. The most productive path forward is to deploy these

complementary approaches in complementary fashion, for the relevant distinction is not

between qualitative and quantitative methods, but rather between tracing the assembly of an

artifact and comparing artifacts once assembled. By these means we may come to see the maps

of social media for what they are: incomplete and unnatural in any given configuration, yet

indispensable in function for navigating the unfathomable largeness of the networked world.

APPENDIX

A.1: The Digg Patriots Data: Origins, Controversies, Considerations

I was first introduced to the Digg Patriots by Drew Harry, who pointed me to

OleOleOlson’s 2010 Alternet article exposing their existence.165 Additional research led me to

an article by Olson published in the NewsJunkiePost a few weeks after the original report.166 In

this article, Olson revealed that the source material for his stories had come from an archive of

the Digg Patriots’ Yahoo!Group listserv. He wrote:

Our investigation team decided early on that we should not release the full archive on Wikileaks due to the presence of personal information on DP members, and they have manipulated this decision to attempt to claim that this is manufactured. The Digg Patriots have been asked numerous times on public forums to agree to sign a release of all mass communications archived from the Yahoo group if they would like to

66

165 Olson, “Massive Censorship of Digg Uncovered.” 166 Olson, “Digg Patriots Censorship Part 2: The Evidence”, NewsJunkiePost, October 15, 2010, http://newsjunkiepost.com/2010/10/15/digg-patriots-censorship-part-2-the-evidence/.

our investigation team to take a much longer approach of redacting personal information manually. That archive is now available at NoTruePatriot.org.

Though there were, by Olson’s count, between 30,000-40,000 messages exchanged

between the Patriots, NoTruePatriot.org contained around 13,000 posts in HTML form, which

Olson attributed to an aggressive redaction program. I was able to access NoTruePatriot.org

from summer 2011 through February 2012. By April 2012, however, the site had been replaced

by a generic landing page for GoDaddy.com, presumably the domain’s registrar. I contacted

Olson and another member of his investigation team, but before they responded to me, the site

was restored. I used a free Firefox extension called DownThemAll to download all of the files

for local storage in case the archive disappeared again. Sure enough, within a few months it

was gone, apparently purchased by domain squatters. To the best of my knowledge there is no

copy of the NoTruePatriot archive now available publicly through the web.

I decided to parse the messages in the HTML files into plain text to assist in my

qualitative analysis. Beth Hadley ’15 helped me write a Python program using the

beautifulSoup library which pulled some important information (date, time, subject, and body)

out of the posts and wrote them to text files. The program also produced a spreadsheet which

(among other things) counted links to Digg.com to help me quantify their contents. This

process yielded 12,742 posts. I imported the text files, along with the original HTML files, into

a DEVONthink database, and then read every post, tagging and highlighting them as I went.

After several days I had tagged several hundred posts with dozens of categories, and I

excerpted from these posts in Chapter II. I have by no means told “the whole story” of the

Patriots. There were many fascinating characters, dynamics, subplots, and so forth which I

67

abridged. Instead, I have tried to tell a faithfully representative story of the Patriots for the

purposes of understanding user-generated censorship as a mode of action.

I have struggled with the question of whether to publish my copy of the NoTruePatriot

archive along with my thesis. On the one hand, it is a vibrant, interesting, rich data set which is

no longer publicly available. Publishing it for general access would both enhance my own work

(by making its source material available alongside it) and also potentially provide opportunities

for future scholarship. On the other hand, the Patriots’ corpus contains a great deal of personal

information, which its members clearly intended to not be available to a general audience.

Although Olson’s “investigative team” did redact most email addresses and names from the

posts, they did not remove all of them, and many of the Patriots can be quickly reidentified

through cursory Google searches.

A major complicating factor is the considerable controversy over how Olson’s team

took possession of the corpus. Olson told me via email that the messages had been leaked by a

“disaffected Digg Patriot.” He pointed me to a post on NewsJunkiePost written under the name

Sam Pennington - according to Olson the investigative leader of the team - which alleged that

the messages had been “forwarded” to them by at least two members of the Patriots.167

Pennington claimed that these members had become disturbed by the online activity of a

Patriot called R.J. Carter. Carter, however, responded by arguing that screenshots posted by

Pennington showed they had been taken by someone logged in as him, and accused

Pennington of hacking his Yahoo account to access the archives.168 However, after the Digg

68

167 Sam Pennington, “The Digg Patriots Investigation: How We Exposed DiggGate”, NewsJunkiePost, October 19, 2010, http://newsjunkiepost.com/2010/10/19/the-digg-patriots-investigation-how-we-exposed-digggate/. 168 R.J. Carter, “NewsJunkiePost Implicated in Computer Hacking to Breach DiggPatriots Yahoo Group,” R.J. Carter’s Line in the Sand, December 6, 2010, http://therjcarter.wordpress.com/2010/12/06/newsjunkiepost-implicated-in-computer-hacking-to-breach-diggpatriots-yahoo-group/.

Patriots were exposed, they began using Skype chats to communicate, and some of the logs of

these chats were later published in what Carter himself called a definite leak.169 Olson, for his

part, claimed that he had always believed Carter was the mole, and had been accusing

Pennington and others of hacking in order to cover his tracks. There is much interesting

intrigue but almost no certainty or clarity surrounding the provenance of the corpus.

On several occasions I attempted to contact the Patriots to interview them directly. I

managed to identify email addresses and Facebook accounts for several Patriots but most failed

to respond to my inquiries. However, my message was forwarded to the current Digg Patriots

mailing list and R.J. Carter (intrigue!) emailed me to discuss his thoughts on the Patriots. I

asked him how the Patriots would feel about me making their archive public again. He told me

that, while he himself did not care since he was proud of their actions, most other members did

not want them to be published anew. This was the clearest signal I had yet about how the

Patriots felt about the corpus I had come to possess.

I have tried to balance the academic interest in the Patriots’ archive with respect for

their privacy as contextual integrity.170 As a result I have also decided not to make my copy of

the Patriots’ archive publicly available at this time, because of my ethical discomfort with

indiscriminately sharing information which had been intended for a small audience and

subsequently published to a general audience without their consent. However, I plan to

provide the corpus to the MIT archives to preserve it for future scholarship, and researchers

may contact me via email at chris [dot] peterson [at] mit [dot] edu to request a copy.

69

169 Olson, email to author, February 11, 2013. 170 See generally Helen Nissenbaum, Privacy in Context: Technology, Policy, and the Integrity of Social Life (Stanford Law Books, 2009).

A.2: The Digg Patriots Data: Some Exploratory Visualizations

I used the spreadsheet which I generated from the Patriots’ corpus to produce some

data visualizations with Tableau. I did not include them in my case study chapter because they

were not ‘findings’ I wanted to share, but rather instruments which guided my exploration of the

corpus, drawing my attention to particular users, patterns of use, and so forth.171

Figure 5: Gantt Projection of the Activity Index

70

171 For more on visualizations-as--exploration, see generally Schnapp et al, Digital_Humanities (MIT Press, 2012).

Figure 5 depicts a Gantt chart of a measure I developed called the “activity index,”

which is the sum of the count of messages sent by a user and the count of links to Digg

submitted by that user. Each dot represents a week during which that user sent a message to

the list, the size of the dot scales with the size of that week’s activity index, and the users are

ranked in order of their activity index over the length of the corpus. This chart illustrated

different patterns of use. Some I expected: for example, that Lizbett was the most active, while

many members barely registered a trace. Some surprised me: minarchian, a libertarian activist

who joined the Patriots latest among its members, was still so active during his admittedly short

tenure that he had one of the highest activity indices overall. Had I been interested in telling

the story of the Digg Patriots qua the Digg Patriots, as opposed to qua user-generated

censorship, a chart like this would have been invaluable in helping me identify a ‘central cast’

on whom to focus my narrative attention. I have included below another Gantt chart which

narrows its scope to a dozen such members to illustrate the point.

Figure 6: A Gantt Projection of the Activity Index for 12 Core Members of the Digg Patriots

71

Figures 7 and 8 flip the Gantt projection on its axis and split the activity index into its

constitutive elements: posts which contain links to Digg, and posts which contain no links to

Digg, which I have called commentary. While the distributions follow the trend of the activity

Figure 7: Distribution of Digg Links Submitted

Figure 8: Distribution of Commentary

72

index, the delineation between links and commentary allowed me to distinguish between types

of users. Notice, for example, that Brendan M. sent far fewer messages with links to Digg than

he did without. A close inspection of his posts reveals that, rather than initiating a course of

action, Brendan preferred to join the fray, and sent lots of gossipy, joking, ‘smalltalk’ messages

back and forth among his friends in the group. A similar pattern can be seen in VRayZ, who,

despite being the second most active user overall behind Lizbett, skewed heavily towards

commentary, which corroborates his status as a senior ‘thought leader’ within the Patriots.

Figure 9: Post Contents Over Time

73

Figures 9 and 10 depict the contents of the posts in the corpus over time. In Figure 9,

the total height of both the blue and orange bars is equal to the number of posts sent to the

group during that week.172 This chart shows that in most weeks the Patriots made more posts

that did not contain links to Digg (blue) than did (orange). However, as Figure 10 shows, the

total number of links per week often exceeded the number of posts sent during that time,

because posts which did contain links contained large numbers of them. These charts illustrate

two modes of engagement with the listserv: the majority of messages which constituted of

commentary, and a minority of messages which were densely packed with links to target for

intervention.

That’s it. Thanks!

Figure 10: Links vs Posts Over Time

74

172 The spike in June and July appears to be an artifact of the redaction process, which was not fully completed for these months according to Olson, rather than an actual increase in activity during that time. This discrepancy itself was revealed in the process of making these visualizations.